Cooperative Relationship Between Pharmaceutical

VOLUME 26 䡠 NUMBER 35 䡠 DECEMBER 10 2008

JOURNAL OF CLINICAL ONCOLOGY

Cooperative Relationship Between Pharmaceutical Companies, Academia, and Media Explains Sharp Decrease in Frequency of Pulmonary Complications After Bortezomib in Japan TO THE EDITOR: Bortezomib (Velcade; Millenium Pharmaceuticals, Cambridge, MA) is a promising new drug for multiple myeloma that was approved by the US Food and Drug Administration in May 2003. Japan lags behind Europe and the United States in the release of new drugs. There, unapproved, privately imported bortezomib was administrated without clinical trials until December 2006, when it was approved and released.1,2 In October 2005, it became clear that patients who were administered bortezomib before approval in Japan had fatal pulmonary complications.3 This was 14 months before bortezomib received approval from the Ministry of Health, Labor and Welfare and appeared on the open market in Japan. At that time, there were few reports about severe pulmonary complications with bortezomib in Europe and the United States.4,5 Medical staff did not recognize that bortezomib might be causing severe pulmonary complications in some patients. By reading reports, medical staff involved in caring for these patients realized the reason for these complications and took measures to reduce the adverse effects. In consequence, it became clear that the frequency of pulmonary complications that occurred before bortezomib became commercially available was dramatically decreased when these adverse effects were reported in the interim report of the postmarketing clinical trial that had been recently disclosed publicly. Information about how the adverse effects of bortezomib were overcome in a short period will prove to be useful in assessing adverse effects of other new drugs. But in the case in question, additional details about the process are still not known. To investigate the frequency of and reasons for pulmonary complications, we reviewed scientific articlespapers, Web-based information from societies and pharmaceutical companies, newspaper reports, and assessment reports from the Pharmaceuticals and Medical Devices Agency (PMDA). We examined information dating from October 2003, when pulmonary complications were first discovered, to December 2007, when the intermediate results of postmarketing clinical tests were reported. We also interviewed the concerned parties (Fig 1). Clinical trials of bortezomib as the sole treatment for multiple myeloma patients were started in Japan by Janssen Pharmaceutical Corp in May 2004. Registration of patients was stopped because one patient died as a result of pulmonary complications after bortezomib administration. As a result of meetings held by a third-party evaluation committee, patients with lung problems seen on imaging tests such as a computed tomography were excluded from the clinical trial. Registration of patients was then resumed.6 Finally, thirty-four patients were registered at the clinical trial; only one (3%) had pulmonary complications. Corticosteroids were not administered in 5820

© 2008 by American Society of Clinical Oncology

C O R R E S P O N D E N C E

Year

Disclosure by academic society

2003

Announcement by PMDA or JPC

Announcement by other media

Approval in US (May 2003)

2004 Clinical trial in Japan (May 2008)

2005 First report to PMDA and JSCH (Oct 2005) First announcement on of the home page of JPC (Oct 2005) First announcement by JSCH (Dec 2005)

2006

Asahi Shinbun (Dec 2005)

Publication of the case series in Blood (Jan 2006)

Nikkei Biotech (Dec 2005)

Interview on MRIC (Feb 2006)

Publication of the nationwide survey in IJH and JSCH (Mar 2006) Approval in Japan (Dec 2006)

2007

Fatal lung injury in a patient with PS 4 (Jan 2007) Recommendation by JPC Announcement of postmarketing surveillance study (Dec 2007)

Fig 1. Disclosure of the information on lung injury after bortezomib in Japanese patients. PMDA, Pharmaceutical and Medical Devise Agency; JPC, Janssen Pharmaceutical Corp; JSCH, Japanese Society of Clinical Hematology; IJH, International Journal of Hematology; JJCH, Japanese Journal of Clinical Hematology (Rinshoketsueki); PS, performance status.

combination with bortezomib to any of the patients. Almost all the patients whose Karnofsky performance status was ⬎ 80% were in better health7 compared with patients who received privately imported bortezomib (Table 1). This information had not been reported by the media until it was made available on Janssen’s Web site in October 2005. In October 2005, Miyakoshi et al suspected that bortezomib was the cause of the pulmonary complications because of their experience administering bortezomib to patients who later suffered from pulmonary complications. Miyakoshi contacted colleagues and researchers and learned that similar cases often occurred (S. Miyakoshi, personal communication, June 2008). Miyakoshi’s research group investigated medical records from 13 patients who were administered bortezomib that was imported between January 2004 and September 2005.3 Four patients (31%) had severe pulmonary complications and two of them (14%) died (Table 1). Four patients who had pulmonary complications were not given corticosteroids at the same time. Performance Journal of Clinical Oncology, Vol 26, No 35 (December 10), 2008: pp 5820-5830

Downloaded from jco.ascopubs.org on October 1, 2012. For personal use only. No other uses without permission. Copyright © 2008 American Society of Clinical Oncology. All rights reserved.

Correspondence

Table 1. Reports of Severe Pulmonary Complications after Bortezomib Treatment Reports

Parameter

Premarketing Clinical Trial by Janssen Pharmaceutical Corp

Study periods

May 2004 to March 2006

Date of public announcement

October 14, 2005 (via Internet)

No. of enrolled patients Karnofsky performance status

34 All the patients are ⱖ 60 (97% of the patients are ⱖ 80)

Median Range Patients who received concomitant corticosteroid treatment No.

% Patients who developed pulmonary complication No. % Patients with fatal pulmonary complication No. %

Clinical Report by Miyakoshi et al June 2004 and September 2005 January 12, 2006 (as Blood First Edition article) 13 NA

Nationwide Surveillance by the Japanese Society of Haematology and the Japanese Society of Clinical Haematologyⴱ June 2003 to July 2005 March 20, 2006 (via Internet and the official journals) 46

Postmarketing Surveillance Study (Preliminary Results) by Janssen Pharmaceutical Corp December 2006 to November 2007 December 26, 2007 (via Internet)

666 NA

70 20-100

0

NA (4 patients with pulmonary complication did not receive steroids)

0

25

124

54

70

1 3

4 31

7 15

24 3.6

1 3

2 14

3 7

4 0.5

Abbreviation: NA, not available. ⴱ Included the 13 patients who had been reported by Miyakoshi et al. †One hundred seventy-eight patients were evaluable.

status (PS; Eastern Cooperative Oncology Group) in these four patients and clinical information for nine patients who had no pulmonary complications were not available. In October 2005, Miyakoshi et al sent the first report about the possibility of pulmonary complications with bortezomib to PMDA, Janssen, and a company that imports bortezomib (RHC USA Corp, Honolulu, HI). PMDA asked Jansses to issue announcements to patients who took part in the clinical trial and to medical staff about the severe pulmonary complications with bortezomib. In the past, information about adverse effects in patients who took part in clinical trials had not been made available except to the hospital that took part in the clinical trial. This time, Jansses showed this information on its Web site on October 24th, 2007, in compliance with PMDA’s request. RHC USA Corp wrote letters to doctors on October 25th, 2005, providing them with the information. Miyakoshi et al. submitted this result to Blood on November 16th, 2005, and posted it online on January 12th, 2006.3 Miyakoshi pointed out the problem of information disclosure related to private medicine importing during an online interview.8 In August and November 2006, American hospitals reported that African Americans had the same severe pulmonary complications with bortezomib.9,10 The Japanese Society of Hematology and the Japanese Society of Clinical Hematology together surveyed self-reported pulmonary www.jco.org

complications on a nationwide scale in November 2005. They showed the results on their Web sites on December 6th, 2005. The results were reported in the Asahi Shimbun newspaper on December 11th, and in Nikkei Biotech on December 19th. Final results of the investigation were shown on both societies’ Web sites on March 20th, 2006, and were also shown in both societies’ academic journals.1,2 A total of 46 patients were administered imported bortezomib from January 2003 to July 2005. Their median Karnofsky PS was 70% (range, 20% to 100%).11 Seven (15%) had pulmonary complications, and three (6.5%) died. Twenty-five (54%) were administered corticosteroids with bortezomib and experienced significantly decreased frequency of pulmonary complications.1 Bortezomib was approved by the Ministry of Health, Labor and Welfare on October 2006, and sales began on the open market in December 2006. After being requested to investigate usage results for all patients, Janssen set a proper usage criterion and recommended that doctors withhold administration of bortezomib from patients whose Karnofsky PS was ⬍ 60 and who had interstitial lesions on their chest radiographs. It came to light in January 2007 that two patients with PS 4 who were administered bortezomib had died of pulmonary complications.12 This was discussed at a third-party evaluation committee headed by Jansses, after which the corporation strongly recommended that doctors not use bortezomib for PS 3 or PS 4 © 2008 by American Society of Clinical Oncology


5821

Correspondence

patients. Intermediate results of postmarketing clinical tests performed in 666 patients who were administered bortezomib after it came to market in November 2007 were reported on Janssen’s Web site in December 2007. Twenty-four (3.6%) had pulmonary complications, and three (0.5%) died (Table 1).13 Information about PS has not been made available to the public. It came into the open in a May 2008 Janssen lecture delivered in Tokyo to commemorate the sale of Velcade that corticosteroids were administered with bortezomib to almost 70% of patients. The drug manufacturer in the United States showed pulmonary complications for Asians in the product document of bortezomib in May 2006.14 Compared with the period when bortezomib was being imported by doctors, the frequency and lethality rate of pulmonary complications at the postmarketing clinical trial decreased. We think some possible causes for this are as follows: first, combined administration with corticosteroids became popular. Fifty-four percent of patients who were administered imported bortezomib were also administered a steroid in combination.3,6 This frequency increased to 70% at the postmarketing clinical trial. The steroid has an antitumor effect on multiple myeloma and is used with many kinds of medicine. Gotoh et al1,2 showed that the risk of pulmonary complications decreased with a combination of bortezomib and the steroid. As a result of this study, some doctors might begin using bortezomib in combination with the steroid for preventing pulmonary complications. Second, as reports of pulmonary complications with bortezomib increased, some doctors began to choose patients carefully. During the clinical trial, Miyakoshi et al found that pulmonary complications occurred in succession with patients in poor general condition who were administered an unapproved drug. After bortezomib went on the market, Janssen set a proper usage criterion that included lung disease history and Karnofsky PS, and the company thoroughly educated doctors. This probably contributed widely to the decrease in adverse events in postmarketing clinical trials. In this case, the period between the discovery of pulmonary complications with bortezomib and the adoption of countermeasures was short. It took only 14 months from the discovery of possible pulmonary complications with bortezomib by Miyakoshi et al to the start of the postmarketing clinical test. During this period, strong measures to reduce the pulmonary complications were taken in clinical practice, as presented in this article. The case highlights some interesting points regarding the communication of medical information. Miyakoshi, who first realized the complications, called and sent e-mails directly to neighborhood hospitals to inquire about this. This shows that doctors can exchange information easily because of the widespread use of information technology. It would have taken much longer for doctors to realize the problem of pulmonary complications with bortezomib without the advantage of information technology. In addition, Miyakoshi et al promptly communicated the risk of adverse effects to the PMDA, Janssen, academic societies, and import agencies. In Japan, a system for collecting information on adverse events related to unapproved drugs has not been established.15 In this case, however, the information was transmitted to doctors comprehensively and promptly because Miyakoshi et al shared their finding of adverse effects related to bortezomib with several organizations, each of which disclosed the information in their own way. Various media reported this problem at an early stage. Mass media such as the Asahi Shinbun newspaper reported that not only medical staff but also patients and their families recognized the adverse 5822


effects of bortezomib. When we think about broadcasting information to the nation, it is necessary to work together with the mass media and medical community, and we should consider the special qualities of each form of media in order to provide information effectively and efficiently. It has been shown in previous situations that mass media such as newspapers can distribute information to many people, but there is a possibility that the information may be incomplete or inaccurate.16 However, although they are targeted only at medical professionals, industry publications such as Medical Research Information Center and Nikkei Biotech can provide more accurate and detailed information in a way that cuts across disciplinary divisions in the world of specialized medicine. In this case, different types of media supplied information concerning pulmonary complications with bortezomib through journals and academic societies, and the risk became widely known in a very short time. The information about pulmonary complications with bortezomib was shared with concerned parties in a short time, and administration of bortezomib was optimized. In this process, in addition to academic societies and pharmaceutical companies, the mass media, use of e-mail, and the Internet played a stronger role.

Hiroto Narimatsu Division of Strategic Outcome Research Program for Cancer Control, Ministry of Health, Labor and Welfare Commission, Japan Cancer Society, Tokyo, Japan; Division of Exploratory Research, the Institute of Medical Science, the University of Tokyo, Tokyo, Japan

Akiko Hori Division of Medical Oncology, Department of Internal Medicine, Teikyo University School of Medicine, Tokyo, Japan

Tomoko Matsumura, Yuko Kodama, Morihito Takita, Yukiko Kishi, Tamae Hamaki, Koichiro Yuji, Yuji Tanaka Division of Exploratory Research, the Institute of Medical Science, the University of Tokyo, Tokyo, Japan

Tsunehiko Komatsu Third Department of Internal Medicine, Teikyo University School of Medicine, Tokyo, Japan

Masahiro Kami Division of Exploratory Research, the Institute of Medical Science, University of Tokyo, Tokyo, Japan

ACKNOWLEDGMENT

This study is supported by a grant from Health and Labor Sciences Research Grants, the third-term comprehensive control for cancer (H18-011) from the Ministry of Health, Labor and Welfare, Tokyo, Japan AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest. REFERENCES 1. Gotoh A, Ohyashiki K, Oshimi K, et al: Lung injury associated with bortezomib therapy in relapsed/refractory multiple myeloma in Japan: A questionnaire-based report from the “Lung Injury by Bortezomib” joint committee of the Japanese society of hematology and the Japanese society of clinical hematology. Int J Hematol 84:406-412, 2006 2. Gotoh A, Ohyashiki K, Oshimi K, et al: Lung injury associated with bortezomib therapy in relapsed/refractory multiple myeloma in Japan: A questionnaire-based report from the “Lung Injury by Bortezomib” Joint Committee of the Japanese Society of Hematology and the Japanese Society of Clinical Hematology [in Japanese]. Rinsho Ketsueki 47:1521-1527, 2006 JOURNAL OF CLINICAL ONCOLOGY


Correspondence

3. Miyakoshi S, Kami M, Yuji K, et al: Severe pulmonary complications in Japanese patients after bortezomib treatment for refractory multiple myeloma. Blood 107:3492-3494, 2006 4. Richardson PG, Barlogie B, Berenson J, et al: A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med 348:2609-2617, 2003 5. Jagannath S, Barlogie B, Berenson J, et al: A phase 2 study of two doses of bortezomib in relapsed or refractory myeloma. Br J Haematol 127:165-172, 2004 6. Pharmaceuticals and Medical Devices Agency: Assessment reports of Bortezomib [in Japanese]. http://www.info.pmda.go.jp/shinyaku/g061004/ 800155000_21800AMX10868000_A100_3.pdf 7. Ogawa Y, Tobinai K, Ogura M, et al: Phase I and II pharmacokinetic and pharmacodynamic study of the proteasome inhibitor bortezomib in Japanese patients with relapsed or refractory multiple myeloma. Cancer Sci 99:140-144, 2008 8. MRIC: MRIC Extraordinary, Vol. 5: Interview with Dr. Miyakoshi [in Japanese]. http://mric.tanaka.md/2006/02/28/2006_mric_vol.html 9. Ohri A, Arena FP: Severe pulmonary complications in African-American patient after bortezomib therapy. Am J Ther 13:553-555, 2006 10. Boyer JE, Batra RB, Ascensao JL, et al: Severe pulmonary complication after bortezomib treatment for multiple myeloma. Blood 108:1113, 2006

11. Japanese Society of Hematology: (Home page) [in Japanese]. http:// www.jshem.or.jp/en/ 12. Janssen Pharmaceutical Corporation: Report from third-party evaluation committee on pulmonary complication [in Japanese]. http://www.janssen.co .jp/sickness/velcade/pdf/discussion-result_070119.pdf 13. Janssen Pharmaceutical Corporation: Intermediate results of postmarketing clinical tests: Bortezomib (December 2007) [in Japanese].http://www .janssen.co.jp/inforest/portalexternal/diviewitemeventdatafiledownload?paf_gear_ id⫽2200032&cid⫽cnt45112 14. US Food and Drug Administration: VELCADE (bortezomib) for injection. http://www.fda.gov/medwatch/safety/2006/Dec_PIs/Velcade_PI.pdf 15. Miyakoshi S, Kusumi E, Kodama Y, et al: Adverse events of unapproved drugs in Japan. Lancet Oncol 9:414-415, 2008 16. Moynihan R, Bero L, Ross-Degnan D, et al: Coverage by the news media of the benefits and risks of medications. N Engl J Med 342:1645-1650, 2000

DOI: 10.1200/JCO.2008.20.3307; published online ahead of print at www.jco.org on November 10, 2008

■ ■ ■

Pleomorphic Lobular Carcinoma of the Breast: Four Long-Term Responders to Trastuzumab—Coincidence or Hint? TO THE EDITOR: Positivity for human growth factor receptor 2 (HER-2) occurs in 20% to 25% of all invasive breast cancers. Infiltrating lobular carcinomas—another relatively uncommon entity— constitute only 7% to 10% of all invasive breast cancers. The pleomorphic lobular carcinoma (PLC) variant accounts for approximately 15% of all invasive lobular carcinomas.1,2 In general, lobular carcinomas are rarely HER-2 positive (8% per Gonzalez-Angulo et al3). In contrast, at least 30% of instances of PLC are identified HER-2 positive by fluorescence in situ hybridization (FISH).4,5 Thus, HER-2–positive PLC should constitute an extremely rare subtype of invasive breast cancers. We describe four patients with PLC who appear to have exquisite sensitivity to trastuzumab. The first patient had a mastectomy for a 7-cm, node-positive, HER-2–positive by FISH, hormone receptor– positive tumor. She was treated with adjuvant chemotherapy and hormonal therapy, but developed biopsy-proved liver metastases 3 years later. She continues to receive trastuzumab at 9 years, with no evidence of recurrent disease. The second patient presented with severe inflammatory carcinoma, and was treated with paclitaxel and trastuzumab, with complete pathologic response in the breast and lymph nodes. She completed 1 year of adjuvant trastuzmab, along with radiation therapy to the left chest wall and draining lymphatics, and has been disease free for 8 years. She did develop a new early-stage primary lung cancer that was treated surgically. The third patient had a primary breast cancer that was treated with mastectomy. Bone metastases were diagnosed shortly after completion of adjuvant chemotherapy. She was treated with capecitabine and trastuzumab for 1.5 years. The capecitabine was then discontinued, and her disease has been totally controlled with the administration of trastuzumab alone for 9 years, with bone imaging demonstrating a few residual sclerotic lesions. The fourth patient had a mastectomy for a 4-cm, nodenegative, HER-2–positive by FISH, hormone receptor–positive tumor characterized as a grade 2 infiltrating lobular carcinoma. According to our local pathologist, this tumor may have been termed a PLC with an www.jco.org

older classification system. She was given adjuvant chemotherapy, and biopsy-proved bone metastases were diagnosed 2 years later. Treatment included vinorelbine and trastuzumab for 1 month, followed by single-agent trastuzumab for 4 years, with no evidence of metabolically active metastatic disease on most recent imaging. Although these four patients may represent a remarkable coincidence, they may instead provide a hint of extreme sensitivity to trastuzumab within this rare subset of patients. We present these patients to suggest that others seek similar associations, which, if present, could stimulate gene microarray studies aimed at finding commonalities within this small group of patients.

Reshma L. Mahtani and Charles L. Vogel Lynn Cancer Institute West Campus, Boca Raton Community Hospital, Boca Raton, FL

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: Reshma L. Mahtani, AstraZeneca (C); Charles L. Vogel, AstraZeneca (C) Consultant or Advisory Role: Charles L. Vogel, Genentech Inc (C), Bristol-Myers Squibb Co (C), Pfizer (C) Stock Ownership: None Honoraria: Charles L. Vogel, GlaxoSmithKline, Bristol-Myers Squibb Co, Genentech Inc Research Funding: Charles L. Vogel, Genentech Inc, Amgen, Pfizer Expert Testimony: None Other Remuneration: None REFERENCES 1. Ferlicot S, Vincent-Salomon A, Medioni J, et al: Wide metastatic spreading in infiltrating lobular carcinoma of the breast. Eur J Cancer 40:336-341, 2004 2. Fadare O, Dadmanesh F, Alvarado-Cabrero I, et al: Lobular intraepithelial neoplasia [lobular carncinoma in situ] with comedo-type necrosis: A clinicopathologic study of 18 cases. Am J Surg Pathol 30:1445-1453, 2006 3. Gonzalez-Angulo AM, Sahin A, Krishnamurthy S, et al: 0 Biologic markers in axillary node-negative breast cancer: Differential expression in invasive



5823

Correspondence

ductal carcinoma versus invasive lobular carcinoma. Clin Breast Cancer 7:396-400, 2006 4. Varga Z, Zhao J, Ohlschlegel C, et al: Preferential HER-2/neu overexpression and/or amplification in aggressive histological subtypes of invasive breast cancer. Histopathology 44:332-338, 2004

5. Turashvili G, Bouchalova K, Bouchal J, et al: Expression of E-cadherin and c-erbB-2/HER-2/neu oncoprotein in high-grade breast cancer. Cesk Patol 43:87-92, 2007


■ ■ ■

Does ER-␤cx Really Have No Clinical Importance in Tamoxifen-Treated Breast Cancer Patients? TO THE EDITOR: We read with interest the article by Honma et al1 on the utility of estrogen receptor ␤ (ER-␤) in breast cancer patients treated with adjuvant tamoxifen, but there are a number of issues we wish to raise with regard to the ER-␤cx data that call into question the authors’ conclusion regarding its use. First, a 10% cutoff was used to determine positivity for ER-␤cx; however this cutoff, although based on published data for ER-␤1, is not the case for ER-␤cx, and no evidence or data were provided to support or validate its use. Previous immunohistochemical studies of ER-␤cx in breast cancer have used either (1) the Allred scoring system, in which samples were defined as being either negative (Allred score 0 to 1) or positive (Allred score 2 to 8)2 or has having low (Allred score 0 to 5) or high ER-␤cx (Allred score 6 to 8) expression,3 or (2) a method in which the exact percentage of cells with nuclear staining was recorded, with staining of 30% or more taken as positive.4 Only one of these studies3 used the same antibody used by Honma et al,1 and within this study, the cutoff for ER-␤cx was defined in a nonarbitrary manner, using a receiver operator curve analysis. It therefore appears that the decision to use a 10% cutoff was arbitrary and not based on any validated data. Given this, the subsequent results and interpretation with regard to ER-␤cx have to be questioned, as others have clearly shown that determination of what constitutes positive or negative within the context of ER␤ staining requires a well validated scoring system alongside correlation with defined clinical outcomes.5 Furthermore, given what is known about the molecular biology and function of ER-␤cx, it is likely that a more detailed analysis of outcome in ER␣-positive tamoxifen-treated patients would have been appropriate. It is known that ER-␤cx preferentially forms heterodimers with ER␣ and ERß; whereas ER␣ is inhibited by ER-ßcx, ER-ß1 is unaffected. Therefore, ER-ßcx dimerization with ER␣ negatively modulates its ligand-binding activity and acts as a dominantnegative inhibitor of ER␣.6 It is therefore possible that the expression of ER-ßcx could block estrogen-dependent signaling via ER␣ and potentially synergize with antiestrogens, such as tamoxifen, thus making endocrine therapy more effective. If this hypothesis is correct, expression of ERßcx could be a useful biomarker in predicting benefit to tamoxifen. To date, all studies that have specifically investigated the possible role of ER-ßcx in modulating response to tamoxifen have been small, and the results have been inconsistent. A study of 50 patients with tamoxifen-resistant tumors (n ⫽ 16) or tamoxifensensitive tumors (n ⫽ 34) reported the tamoxifen-sensitive group to have a higher positivity for ER-ßcx.4 A study of 23 patients with assessable disease found ER-ßcx expression to be significantly associated with a response to endocrine therapy and with longer survival.7 In 56 patients given adjuvant tamoxifen alone, high ER-ßcx mRNA levels

were significantly associated with an improvement in overall survival.8 High ER-ßcx mRNA levels, but not protein levels, were independently predictive of outcome in 85 patients treated with adjuvant tamoxifen,3 and a neoadjuvant study of 18 women found ER-ßcx to be associated with a poor clinical response to tamoxifen.2 Therefore, further work is needed to determine the potential role, if any, of ER-ßcx as a biomarker in endocrine therapy. In light of these issues, it would appear essential that the ERßcx immunostaining in the ER␣-positive patients exposed to tamoxifen be rescored using the Allred method and that a receiver operator characteristic curve analysis be performed to select a clinically relevant cutoff.9 The data could then be used to address the utility of ER-ßcx in determining sensitivity or resistance to tamoxifen therapy.

Carlo Palmieri Department of Cancer Medicine, Imperial College London, Cancer Research UK Laboratories, London, United Kingdom

Ondrej Gojis Department of Pathology, University Hospital Kralovske Vinohrady, Third Faculty of Medicine, Charles University, Prague, Czech Republic

Bharath Rudraraju and Susan Cleator Department of Cancer Medicine, Imperial College London, Cancer Research UK Laboratories, London, United Kingdom


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Honma N, Horii R, Iwase T, et al: Clinical importance of estrogen receptor-␤ evaluation in breast cancer patients treated with adjuvant tamoxifen therapy. J Clin Oncol 28:3727-3734, 2008 2. Saji S, Omoto Y, Shimizu C, et al: Expression of estrogen receptor (ER) betacx protein in ERalpha-positive breast cancer: Specific correlation with progesterone receptor. Cancer Res 62:4849-4853, 2002 3. Vinayagam R, Sibson DR, Holcombe C, et al: Association of oestrogen receptor beta 2 (ER beta 2/ER beta cx) with outcome of adjuvant endocrine treatment for primary breast cancer: A retrospective study. BMC Cancer 7:131, 2007 4. Esslimani-Sahla M, Simony-Lafontaine J, Kramar A, et al: Estrogen receptor beta (ERbeta) level but not its ERbetacx variant helps to predict tamoxifen resistance in breast cancer. Clin Cancer Res 10:5769-5776, 2004 5. Carder PJ, Murphy PJ, Dervan P, et al: A multi-centre investigation towards reaching a consensus on the immunohistochemical detection of ERbeta in archival formalin-fixed paraffin embedded human breast tissue. Breast Cancer Res Treat 92:287-293, 2005 6. Ogawa S, Inoue S, Watanabe T, et al: Molecular cloning and characterization of human estrogen receptor betacx: A potential inhibitor of oestrogen action in human. Nucleic Acids Res 26:3505-3512, 1998 7. Palmieri C, Lam EW, Mansi J, et al: The expression of ER beta cx in human breast cancer and the relationship to endocrine therapy and survival. Clin Cancer Res 10:2421-2428, 2004 8. Davies MP, O’Neill PA, Innes H, et al: Correlation of mRNA for oestrogen receptor beta splice variants ERbeta1, ERbeta2/ERbetacx and ERbeta5 with outcome in endocrine-treated breast cancer. J Mol Endocrinol 33:773-782, 2004 9. Hanley J: Receiver operating characteristic (ROC) methodology: The state of the art. Critical Rev Diagn Imagin 29:307-337, 1989


■ ■ ■

5824




Correspondence

Clinical Importance of Estrogen Receptor ␤ Isoforms in Breast Cancer TO THE EDITOR: Since its discovery in 1996,1 progress in understanding the potential role of estrogen receptor ␤ (ER␤) in the clinical management of breast cancer has been hindered by the publication of contradictory data from studies that used small numbers of cases, relied on poorly validated primary antibodies, or failed to account for the contribution of the five known ER␤ isoforms.2,3 Recently, the situation has improved through the publication of two large studies. In Journal of Clinical Oncology, in one of the largest immunohistochemical studies conducted to date, Honma et al4 examined total ER␤, ER␤1, and ER␤2 in 442 breast tumors from a group of Japanese patients who all received tamoxifen monotherapy. They showed that ER␤1 was an independent predictor of recurrence, disease-free survival, and overall survival, particularly in the so-called triple-negative breast cancers, which do not express ER␣, progesterone, or HER2. In our immunohistochemical study, we evaluated expression of ER␤1, ER␤2, and ER␤5 in a cohort of 880 unselected breast tumors of European origin.5 In contrast to the Honma et al study, our study revealed that ER␤2 was the most significant ER␤ isoform in breast cancer, as it was predictive of disease-free and overall survival and response to hormone therapy. Using the same antibody as Honma et al,4 we found that ER␤1 was uninformative, irrespective of scoring method or cutoff used. Our study revealed a moderate association of ER␤5 with improved survival, but only when it was strongly expressed. It is possible that ethnic differences between the two cohorts may have contributed to these contrasting observations, as significant differences in expression of ER␤ isoforms have been reported at the mRNA level in African-American versus Caucasian women.6 A comparison of these cohorts showed that the former group expressed significantly higher levels of ER␤1 and, in particular, ER␤5. No differences were seen with ER␣, and ER␤2 was not examined. These three studies4-6 strongly suggest that the specific type and level of ER␤ isoform expression may contribute to breast cancer survival across different ethnic groups. Finally, the presence of cytoplasmic ER␤ immunoreactivity has been consistently noted in a number of studies, including that of Honma et al.4 We have formally examined this and have shown that

the cellular location of ER␤ isoforms determines outcome. Although nuclear ER␤2 predicted good clinical response, cytoplasmic ER␤2 expression, alone or combined with nuclear staining, predicted significantly worse overall survival. Notably, patients with only cytoplasmic ER␤2 expression had significantly worse outcomes.5 In summary, not only does the type of ER␤ expressed influence survival, but the cellular location of ER␤ isoforms, in particular ER␤2, has distinct prognostic outcome in breast cancer. We suggest that ER␤ isoforms should be considered in translational arms of future breast cancer trials to determine whether they can provide additional clinically useful information to oncologists involved in direct breast cancer treatment.

Valerie Speirs, Andrew R. Green, Thomas A. Hughes, Ian O. Ellis, Philippa T.K. Saunders, and Abeer M. Shaaban Leeds Institute of Molecular Medicine, University of Leeds, Leeds, United Kingdom; Department of Histopathology, Nottingham University Hospitals Trust and University of Nottingham, Nottingham, United Kingdom, Medical Research Council Human Reproductive Sciences Unit, Queen’s Medical Research Institute, Edinburgh, United Kingdom


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Kuiper GG, Enmark E, Pelto-Huikko M, et al: Cloning of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci U S A 93:5925-5930, 1996 2. Speirs V, Carder PJ, Lane S, et al: Oestrogen receptor ␤: What it means for patients with breast cancer. Lancet Oncol 5:174-181, 2004 3. Moore JT, McKee DD, Slentz-Kesler K, et al: Cloning and characterisation of human estrogen receptor ␤ isoforms. Biochem Biophys Res Comm 247:7578, 1988 4. Honma N, Horri R, Iwase T, et al: Clinical importance of estrogen receptor-␤ evaluation in breast cancer patients treated with adjuvant tamoxifen therapy. J Clin Oncol 26:3727-3734, 2008 5. Shaaban AM, Green AR, Karthik S, et al: Nuclear and cytoplasmic expression of ER␤1, -2 and -5 identifies distinct prognostic outcome for breast cancer patients. Clin Cancer Res 14:5228-5235, 2008 6. Poola I, Fuqua SAW, De Witty RL, et al: Estrogen receptor ␣-negative breast cancer tissues express significant levels of estrogen-independent transcription factors, ER␤1 and ER␤5: Potential molecular targets for chemoprevention. Clin Cancer Res 11:7579-7585, 2005


■ ■ ■

IN REPLY: We thank Dr. Palmieri, Dr. Speirs, and their colleagues 1

for their interest in our recent article and for the thoughtful comments, which we are happy to address. We have already scored all of our immunostaining according to the Allred method, but we elected to use 10% as the cutoff in the final analysis because this is the most commonly used cutoff value for both estrogen receptor-␣ (ER-␣; classic ER) and ER-␤.1 In response to the comments by Palmieri et al, we have reanalyzed our data as suggested. Receiver operating characteristics analysis identified an Allred score of 5 or more as the most suitable cutoff value for ER-␤cx (or ER-␤2) positivity in the present study setting. According to this new cutoff value, the log-rank test for survival difference by ER-␤cx status was P ⫽ .025 and P ⫽ .113 for diseasefree survival (DFS) and overall survival (OS), respectively. This is similar to our results with the 10% cutoff (P ⫽ .030 and P ⫽ .038 www.jco.org

for DFS and OS, respectively).1 ER-␤cx positivity seemed to be associated with a favorable clinical outcome when ER-␤cx was the only factor examined; however, when ER-␤1 status and ER-␤N status were also taken into consideration, Bonferroni adjustment had to be performed, and the difference in clinical outcome according to ER-␤cx status became nonsignificant. When multivariate analysis for recurrence or mortality was performed, taking into consideration the status of ER-␤cx, ER-␤1, and ER-␤N, ER-␤1 emerged as the only independent predictor of clinical outcome, suggesting that ER-␤1 has the most predictive power. In addition, unlike ER-␤1, ER-␤cx was not an independent predictor of clinical outcome (P ⫽ .100 and P ⫽ .367 for recurrence and mortality, respectively) in multivariate analysis that included tumor size, nodal status, grade, human epidermal growth factor 2 status, ER-␣ status, and progesterone receptor status. © 2008 by American Society of Clinical Oncology


5825

Correspondence

On the basis of the comments by Palmieri et al, we have repeated survival analysis according to ER-␤cx status, using the new cutoff value for patients with ER-␣–positive tumors who received adjuvant tamoxifen monotherapy. The clinical outcome of these patients did not differ by ER-␤cx status (P ⫽ .226 and P ⫽ .516 for DFS and OS, respectively). Because both tamoxifen and ER-␤cx inhibit the function of ER-␣,2 the effect of ER-␤cx on clinical outcome may have been obscured by tamoxifen. The dominant negative effect of ER-␤cx against ER-␣ suggested by a molecular biology study2 may be clearly demonstrated in patients who did not receive adjuvant tamoxifen. A similar study restricted to tamoxifen-untreated patients will help to reveal the role of ER-␤cx in clinical breast cancers. Contrary to the results of the study by Shaaban et al,3 ER-␤1 nuclear positivity by itself was significantly associated with a favorable clinical outcome irrespective of the cutoff value (Allred score ⱖ 3, ⱖ 4, ⱖ 5, and ⱖ 6) in our study setting. Using appropriate statistical methods, we did not find the predictive value of ER-␤cx for clinical outcome in tamoxifen-treated patients suggested by Shaaban et al,3 as shown above. We were also interested in the cytoplasmic staining of ER-␤s and have already evaluated it according to the Allred method, although the data were not included in our article. Unlike Shaaban et al,3 we found that cytoplasmic staining of ER-␤cx was not associated with clinical outcome, irrespective of the cutoff value. It is difficult to properly compare our results with those of Shaaban et al,3 because of the different study settings. First, regarding patient treatment, all patients in our study received adjuvant tamoxifen monotherapy, whereas the patients in the study by Shaaban et al.3 were “unselected” and probably included hormone receptor-positive patients who did or did not receive adjuvant endocrine treatment and patients who received adjuvant chemotherapy. Second, there are many methodological differences in immunohistochemistry: sections examined (tissue microarrays in the Shaaban et al study 3 v representative slides in ours1), concentration of the primary antibody, incubation time and temperature, cutoff value, and so on. Given these important differences, it is not surprising that Shaaban et al reached a different conclusion. We agree that ethnic differences may influence

which ER-␤ isotypes are frequently expressed in breast cancers, leading to different patient survival. In conclusion, the finding of our study regarding the clinical significance of ER-␤cx expression did not change, even after we used a new cutoff value, as recommended by Palmieri et al, or after we took cytoplasmic staining into consideration, as recommended by Speirs et al. ER-␤cx nuclear status had no significant difference on clinical outcome in patients with ER-␣-positive tumors treated with adjuvant tamoxifen monotherapy. A large systematic study using archival material and uniform immunostaining and evaluation methods for patients from different ethnic groups with predefined adjuvant treatments is needed to confirm the clinical importance of ER-␤ expression in breast cancer.

Naoko Honma Research Team for Geriatric Diseases, Tokyo Metropolitan Institute of Gerontology, Tokyo, Japan and Department of Breast Pathology, Cancer Institute, Tokyo, Japan

Shigehira Saji Division of Clinical Trials and Research, Department of Breast Surgery, Tokyo Metropolitan Cancer and Infectious Diseases Center, Komagome Hospital, Tokyo, Japan

Mamoun Younes Department of Pathology, Baylor College of Medicine, Houston, TX


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Honma N, Horii R, Iwase T, et al: Clinical importance of estrogen receptor-␤ evaluation in breast cancer patients treated with adjuvant tamoxifen therapy. J Clin Oncol 26:3727-3734, 2008 2. Ogawa S, Inoue S, Watanabe T, et al: Molecular cloning and characterization of human estrogen receptor betacx: A potential inhibitor of oestrogen action in human. Nucleic Acids Res 26:3505-3512, 1998 3. Shaaban AM, Green AR, Karthik S, et al: Nuclear and cytoplasmic expression of ER␤1, ER␤2, and ER␤5 identifies distinct prognostic outcome for breast cancer patients. Clin Cancer Res 14:5228-5235, 2008


■ ■ ■

Paradoxical Clinical Effects of Epidermal Growth Factor Receptor– Tyrosine Kinase Inhibitors for Acute Myelogenous Leukemia TO THE EDITOR: We read with great interest the article by Pitini et al1 in which they reported that erlotinib, an epidermal growth factor receptor–tyrosine kinase inhibitor (EGFR-TKI), was effective for combined EGFR-negative acute myelogenous leukemia (AML) –M1 and EGFR mutation–positive non–small-cell lung cancer (NSCLC). Another patient who had both AML-M1 and NSCLC also benefited substantially from the erlotinib treatment.2 These results are supported by in vitro studies in which TKIs have antiproliferative effects against AML.3,4 Contrary to their exciting case presentation, we and others previously observed the emergence of AML-M3 during longterm treatment with gefitinib (250 mg/d), another TKI, in four pa5826


tients with NSCLC.5,6 We find these paradoxical clinical results rather interesting. There are several possible explanations for the results. First, some cytotoxic drugs or radiotherapy that the patient had undergone before the gefitinib treatment, rather than the gefitinib itself, might have been a risk factor for leukemogenesis in our patients. The greater the improvement in patient survival (the actual interval between the initial treatment and the emergence of AML-M3, which was 35, 48, and 69 months in our three patients), the more frequently we might observe secondary malignancies. In addition, each case was diagnosed at an early stage of AML-M3 with mild to moderate cytopenia and no disseminated intravascular coagulation, and leukemic cells could be detected only in the bone marrow in two of the three patients. Based on these clinical features, we postulate that in our three patients, gefitinib suppressed the proliferation of leukemic cells to an extent, obscuring the diagnosis of AML-M3, despite the need for further relevant investigations. As another explanation for the paradoxical findings, TKIs might have different leukemogenic effects among the subtypes of JOURNAL OF CLINICAL ONCOLOGY


Correspondence

AML. We consider further research is required to clarify whether TKIs might induce especially the BCR-ABL fusion gene, the genetic cause of AML-M3.7 We agree with the conclusion of Pitini et al.1 that TKIs are a potential therapeutic strategy for AML. From this perspective, they shed light on an exciting aspect of EGFR-TKIs that needs to be evaluated in future preclinical and clinical evaluations. In addition, worldwide epidemiological research might clarify the association between long-term TKI treatment and the rare events that we have noted.5

Katsuyuki Hotta, Katsuyuki Kiura, and Nagio Takigawa Department of Respiratory Medicine, Okayama University Hospital, Okayama, Japan

Keitaro Matsuo Division of Epidemiology and Prevention, Aichi Cancer Center Research Institute, Nagoya, Japan


The authors indicated no potential conflicts of interest. REFERENCES 1. Pitini V, Arrigo C, Altavilla G: Erlotinib in a patient with acute myelogenous leukemia and concomitant non–small-cell lung cancer. J Clin Oncol 26:36453646, 2008 2. Chan G, Pilichowska M: Complete remission in a patient with acute myelogenous leukemia treated with erlotinib for non–small-cell lung cancer. Blood 110:1079-1080, 2007 3. Stegmaier K, Corsello SM, Ross KN, et al: Gefitinib induces myeloid differentiation of acute myeloid leukemia. Blood 106:2841-2848, 2005 4. Boehrer S, Ade`s L, Braun T, et al: Erlotinib exhibits antineoplastic off-target effects in AML and MDS: A preclinical study. Blood 111:2170-2180, 2008 5. Uchida A, Matsuo K, Tanimoto M: APL during gefitinib treatment for non–small-cell lung cancer. N Engl J Med 352:842-843, 2005 6. Ennishi D, Sezaki N, Senoo T, et al: A case of acute promyelocytic leukemia during gefitinib treatment. Int J Hematol 84:284-285, 2006 7. Cortes J, Kantarjian H: New targeted approaches in chronic myeloid leukemia. J Clin Oncol 23:6316-6324, 2005

Masahiro Tabata, Yoshiro Fujiwara, and Mitsune Tanimoto DOI: 10.1200/JCO.2008.19.5685; published online ahead of print at www.jco.org on November 10, 2008

Department of Respiratory Medicine, Okayama University Hospital, Okayama, Japan

■ ■ ■

IN REPLY: We thank Hotta et al for their interesting comments 1

about our article. Regarding their first point, we agree to the question as to whether tyrosine kinase inhibitors might induce the reciprocal chromosome translocation t(15;17) (q22;q11-12), the genetic cause of acute myeloid leukemia (AML) –M3; it is not known whether cytotoxic drugs or radiotherapy rather than the gefitinib itself was a risk factor for leukemogenesis, even if an experimental study has demonstrated that gefitinib induces myeloid differentiation of AML.2 Furthermore, gefitinib has been used to treat more than 10,000 patients with locally advanced or metastatic non–small-cell lung cancer with no reported deleterious effects on normal hematopoiesis. The suggestion that gefitinib suppressed the proliferation of leukemic cells to an extent, obscuring the diagnosis of AML-M3, is intriguing. At the genetic level, acute promyelocytic leukemia is characterized by a unique reciprocal chromosome translocation leading to a fusion between the promyelocytic (PML) gene on chromosome 15 and the retinoic acid receptor alpha (RARA) gene on chromosome 17. Based on the PML breakpoint location, the PML/ RARA transcript subtypes bcr1, bcr2, and bcr3 (short type) may be formed. Interesting FLT3 receptor mutations including point mutations in the tyrosine kinase II domain have been detected at high frequency in acute promyelocytic leukemia, and these mutations have been associated with higher WBC count at presentation and with the short type PML/RARA fusion.3 Unfortunately, in the four reported

patients with non–small-cell lung cancer and AML-M3,4,5 the PML breakpoint location has not been formed, so no definitive conclusions can be drawn whether gefitinib suppressed the proliferation of leukemic cells.

Vincenzo Pitini, Carmela Arrigo, and Giuseppe Altavilla Department of Medical Oncology, University of Messina, Messina, Italy


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Pitini V, Arrigo C, Altavilla G: Erlotinib in a patient with acute myelogenous leukemia and concomitant non–small-cell lung cancer. J Clin Oncol 26:36453646, 2008 2. Stegmaier K, Corsello SM, Ross KN, et al: Gefitinib induces myeloid differentiation of acute myeloid leukemia. Blood 106:2841-2848, 2005 3. Lo-Coco F, Ammatuna E, Montesinos P, et al: Acute promyelocytic leukemia: Recent advances in diagnosis and management. Semin Oncol 35:401409, 2008 4. Uchida A, Matsuo K, Tanimoto M: APL during gefitinib treatment for non–small-cell lung cancer. N Engl J Med 352:842-843, 2005 5. Ennishi D, Sezaki N, Senoo T, et al: A case of acute promyelocytic leukemia during gefitinib treatment. Int J Hematol 84:284-285, 2006


■ ■ ■

Is Alemtuzumab Really the Single Active Agent for Treatment of Chronic Lymphocytic Leukemia? TO THE EDITOR: We read with great interest the article by Hillmen et al1 on first-line therapy for chronic lymphocytic leukemia (CLL) and would like to commend the efforts of the authors, but we have some queries and suggestions. www.jco.org

First, regarding patient selection and characteristics, two thirds of the patient population who received alemtuzumab had early-stage (stage 1 or 2) disease with no clear information about their symptoms. No bone marrow examination was performed before the treatment and no computed tomography scans were performed before enrollment onto the study. This can lead to understaging of the patients involved in the study and potential overestimation of the reported benefits of treatment. Second, regarding follow-up evaluations, patients were evaluated by physical examination, laboratory tests, and chest x-rays only. This is © 2008 by American Society of Clinical Oncology


5827

Correspondence

most likely considered suboptimal in clinical practice, and can easily lead to overestimation of the benefits of treatment. For example, while assessing disease progression and response, no consideration would be given to organomegaly and abdominal lymphadenopathy unless more imaging studies were performed. Third, regarding outcomes of the different treatments, the results of this study show that alemtuzumab is superior to chlorambucil with regard to progression-free survival (PFS). If we look at the study by Rai et al2 from 2000, PFS for chlorambucil was 14 months versus 20 months for fludarabine, whereas in this study PFS for alemtuzumab was 14.6 months and 11.7 months for chlorambucil. The median time to alternative therapy for alemtuzumab, according to the current study is 23.3 months versus 14.7 months for chlorambucil. In the study by Rai et al, the median time to alternative therapy was 27 months for fludarabine versus 14 months for cholambucil. Although no comparisons should be made across different studies, the reported outcomes seem to favor fludarabine over alemtuzumab. Complete remission rates were reported in 24.2% of patients with alemtuzumab versus 2% with cholambucil. Byrd et al3 reported that 47% of patients with CLL had complete remission with fludarabine and rituxan followed by rituxan alone. This is almost twice the complete response rate reported in the current study. In addition, alemtuzumab is not a benign chemotherapeutic agent, given that 15% of patients had symptomatic cytomegalovirus and 52% of patients had asymptomatic cytomegalovirus infections.

Therefore, the conclusion that alemtuzumab is the most active single agent for the treatment of patients with CLL seems too ambitious, and the current alemtuzumab regimen should be compared with studies using at least fludarabine alone or fludarabine and rituxan as a comparative arm.

Satinderjit S. Oberoi and Rony M. Abou Jawde Department of Hematology-Oncology, Heartland Regional Medical Center, Saint Joseph, MO


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Hillmen P, Skotnicki AB, Robak T, et al: Alemtuzumab compared with chlorambucil as first-line therapy for chronic lymphocytic leukemia. J Clin Oncol 25:5616-5623, 2007 2. Rai KR, Peterson BL, Appelbaum FR, et al: Fludarabine compared with chlorambucil as primary therapy for chronic lymphocytic leukemia. N Engl J Med 343:1750-1757, 2000 3. Byrd JC, Peterson BL, Morrison VA, et al: Randomized phase 2 study of fludarabine with concurrent versus sequential treatment with rituximab in symptomatic, untreated patients with B-cell chronic lymphocytic leukemia: Results from Cancer and Leukemia Group B 9712 (CALGB 9712). Blood 101:6-14, 2003


■ ■ ■

IN REPLY: Unfortunately, the correspondence of Oberoi and Jawde reveals some confusion in the interpretation of the 1996 National Cancer Institute (NCI) criteria for chronic lymphocytic leukemia (CLL)1 and of the literature. First, the observation that only one third of patients in the CAM307 trial had cytopenia (Rai stage III or IV) indicates that this study is similar to other recently published frontline studies, such as the Leukemia Research Fund CLL4 trial,2 in which only 30% of patients were Binet stage C; and the German CLL4 trial,3 in which 40% of patients were Rai stage III or IV. Second, computed tomography (CT) scans are not required for staging of patients or for the assessment of response, according to the NCI criteria. Therefore, to suggest that CT scans should be used to assess the time to progression is not appropriate. This is because CLL is largely a disease affecting the blood and bone marrow, and when there is nodal disease it is usually palpable. There is no evidence that regular CT scans have a major impact on the assessment of the time to progression. Such an approach would also be associated with numerous CT scans, which carries a potential risk to the patient and certainly has resource implications. Also, the analysis of the bone marrow has no significance for staging patients with CLL, and the pattern of infiltration by CLL in the marrow is of no prognostic importance. Thus, the statement, “this can lead to understaging of the patients involved in the study and potentially overestimating of the reported benefits of treatment” is inaccurate and misleading. While Oberoi and Jawde are correct to state that the comparison of responses between trials is not entirely valid, it is a fact that the complete remission (CR) rate for alemtuzumab in our trial (24% by Independent Review Panel, and 30% by the investigator) is higher 5828


than the fludarabine monotherapy CR rate in other large trials (LRF CLL4 fludarabine arm, investigators assessment, CR ⫽ 15%; and German CLL Study Group CLL4 fludarabine arm, investigators assessment, CR ⫽ 6.7%). In addition, the eradication of minimal residual disease is not reported with fludarabine monotherapy, but was seen in 7.4% of patients treated with alemtuzumab. The combination of fludarabine plus rituximab is not a fair comparator as it cannot be considered monotherapy. Oberoi and Jawde have failed to appreciate the principle advantages of alemtuzumab over fludarabine-based treatment, which may indicate where the role of the drug will lie earlier in the patient’s treatment. First, the mechanism of alemtuzumab is p53independent and because p53 deletion and/or mutation is the most common mechanism of resistance in CLL, this is an important feature. Second, alemtuzumab can eradicate detectable minimal residual disease, and this is associated with prolonged remission and improved survival. Therefore, in our opinion, the comparison of alemtuzumab monotherapy to fludarabine alone (proven to be less effective than fludarabine plus cyclophosphamide combination therapy) or to fludarabine plus rituximab (a therapy for which there is no randomized, controlled data to support its use) is inappropriate. Our study in fact demonstrates that alemtuzumab is relatively safe compared with chlorambucil, and we believe that alemtuzumab should be used in a much more intelligent manner than simply as a comparison with another nonstandard therapy.

Peter Hillmen Leeds Teaching Hospitals National Health Service Trust, St James Institute of Oncology, Leeds, United Kingdom JOURNAL OF CLINICAL ONCOLOGY


Correspondence


REFERENCES

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Peter Hillmen, Bayer Healthcare Research Funding: Peter Hillmen, Genzyme, Bayer Healthcare Expert Testimony: None Other Remuneration: None

1. Cheson BD, Bennett JM, Grever M, et al: National Cancer Institutesponsored working group guidelines for chronic lymphocytic leukemia: Revised guidelines for diagnosis and treatment. Blood 87:4990-4997, 1996 2. Catovsky D, Richards S, Matutes E, et al: Assessment of fludarabine plus cyclophosphamide for patients with chronic lymphocytic leukaemia (the LRF CLL4 Trial): A randomised controlled trial. Lancet 370:230-239, 2007 3. Eichhorst BF, Busch R, Hopfinger G, et al: Fludarabine plus cyclophosphamide versus fludarabine alone in first-line therapy of younger patients with chronic lymphocytic leukemia. Blood 107:885-891, 2006


■ ■ ■

Cost Effectiveness of Sunitinib TO THE EDITOR: In comparing an innovative treatment versus a reference treatment, the cost per life year gained makes sense when both of the two following criteria are met: (1) survival is prolonged, and (2) its prolongation is statistically significant. This issue has long been debated in the literature,1 and also on the pages of Journal of Clinical Oncology.2 The authors of the original trial comparing sunitinib versus interferon in metastatic renal cell carcinoma stated that “although there was a trend toward improved survival with sunitinib […] the comparison did not meet the prespecified level of significance for this interim analysis.”3 Likewise, Remák et al4 conducted their cost-effectiveness based on the assumption of noninferiority of sunitinib “based on expert clinical opinion, long-term overall survival with sunitinib could be equal to, but not worse than, that with interferon at any point,” and therefore negated, at least in the beginning of their analysis, the hypothesis that sunitinib offers an advantage in overall survival. It is thus surprising to see that Remák et al4 have not accepted the conclusion of Motzer et al3 (no demonstrated survival advantage for sunitinib) nor have they accepted the opinion of their own panel of experts (noninferiority of sunitinib), but have superficially gone to the conclusion that sunitinib determines a survival advantage (gain ⫽ 0.11 years per patient v interferon) and that the cost per life year gained can therefore be calculated.

Since this superficial methodological choice was in favor of sunitinib, one cannot help noticing that this study was not an independent one,5,6 but was sponsored by the manufacturer of sunitinib.

Sabrina Trippoli and Andrea Messori Servizio Farmaceutico e Laboratorio di Farmacoeconomia, Azienda Ospedaliera Universitaria di Careggi, Firenze, Italy


The author(s) indicated no potential conflicts of interest. REFERENCES 1. Trippoli S, Messori A: Cost-effectiveness analyses of statistically ineffective treatments. JAMA 280:1992-1993, 1998 2. Trippoli S, Vaiani M, Messori A, et al: Survival gain in cost-effectiveness studies. J Clin Oncol 18:3318, 2000 3. Motzer RJ, Hutson TE, Tomczak P, et al: Sunitinib versus interferon alpha in metastatic renal-cell carcinoma. N Engl J Med 356:115-124, 2007 4. Rema´k R, Charbonneau C, Ne´grier S, et al: Economic evaluation of sunitinib malate for the first-line treatment of metastatic renal cell carcinoma. J Clin Oncol 26:3995-4000, 2008 5. Bell CM, Urbach DR, Ray JG, et al: Bias in published cost-effectiveness studies: Systematic review. BMJ 332:699-703, 2006 6. Dobson R: Industry-sponsored studies twice as likely to have positive conclusions about costs. BMJ 327:1006, 2003


■ ■ ■

IN REPLY: We thank Drs Trippoli and Messori for their comments on our publication.1 The authors are correct in noting that the issue of application of statistical methods in economic evaluation has long been debated. However, the conclusion is not as clear as the authors of this letter would have us believe. Lack of evidence of difference between the treatment arms is not evidence of lack of difference. Furthermore, Drs Trippoli and Messori cite an excerpt from the original clinical report of a randomized phase III study published in New England Journal of Medicine2 to suggest that the criterion for evidence of prolonged survival was not met, but this was taken out of context. The primary analysis of the results was based on the progression-free survival, which was the primary end point. The overall survival was the secondary end point, and at the time of the analysis, the medians were not reached in either treatment group and www.jco.org

therefore did not meet the prespecified level of significance; thus, no conclusion could be made regarding survival. However, this has no bearing on the economic evaluation itself. We share the view with many other health economists that “rules of inference are arbitrary and entirely irrelevant to the decisions which clinical and economic evaluations claim to inform.”3 Decisions should be based on the mean net benefit of the treatment. Cost-minimization analysis, which the authors of this letter seem to call for, has been shown to be an appropriate method of analysis only under rare circumstances (eg, a randomized trial has been designed to test the explicit hypothesis of equivalence in outcome between two therapies), and “the analytic focus should be on the estimation of the joint density of cost and effect differences, the quantification of uncertainty surrounding the incremental cost-effectiveness ratio and the presentation © 2008 by American Society of Clinical Oncology


5829

Correspondence

of such data as cost-effectiveness acceptability curves”.4 All of this has been performed, taking into account all statistical properties of the survival curves reported by the clinical trial, and results have been presented accordingly. The noninferiority assumption of long-term survival with sunitinib was applied in the extrapolation of survival, and is actually a conservative assumption that works against sunitinib and is, therefore, not biased against the comparator in this trial. Without this assumption, the projected overall-survival curves would cross at 174 weeks. The majority of patients are projected to be dead by this point in time, and the decrease in mean overall survival compared with the base case presented in the paper is outweighed by the decrease in cost of best supportive care, making sunitinib more effective and less costly than interferon alfa. We feel that the data and analyses performed were presented clearly, following all recommendations on good practices in modeling and reporting of economic evaluations using modeling,5,6 irrespective of affiliation of the authors. The authorship reflects contribution to the paper, and rightly “JCO makes its editorial decisions based on scientific quality and importance, not authorship.”7

Edit Rema´k Health Care Analytics, United Biosource Corp, London, United Kingdom

AUTHOR’S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest. REFERENCES 1. Rema´k E, Charbonneau C, Ne´grier S, et al: Economic evaluation of sunitinib malate for the first-line treatment of metastatic renal cell carcinoma. J Clin Oncol 26:3995-4000, 2008 2. Motzer RJ, Hutson TE, Tomczak P, et al: Sunitinib versus interferon alfa in metastatic renal-cell carcinoma. N Engl J Med 356:115-124, 2007 3. Claxton K: The irrelevance of inference: A decision-making approach to the stochastic evaluation of health care technologies. J Health Econ 18:341-364, 1999 4. Briggs AH, O’Brien BJ: The death of cost-minimization analysis? Health Econ 10:179-184, 2001 5. Consensus Conference in Guidelines on Economic Modelling in Health Technology Assessment: Decision analytic modelling in the economic evaluation of health technologies. Pharmacoeconomics 17:443-444, 2000 6. Nuijten MJC, Pronk MH, Brorens MJA, et al: Reporting format for economic evaluation. Part II: Focus on modelling studies. Pharmacoeconomics 14:259-268, 1998 7. Levine MN. In reply. J Clin Oncol 25:614-616, 2007


■ ■ ■

5830


