|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 28 (October 1), 2007: pp. 4504-4505 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.12.3901
In ReplyDivisions of Surgery and Information Sciences City of Hope National Medical Center, Duarte, CA Houseman et al express a number of concerns regarding the analyses and conclusions of our recent publication on Merkel cell carcinoma (MCC).1 Their criticism focuses on a number of important issues and purports to raise questions regarding the primary conclusion that adjuvant radiation therapy improves the survival of patients undergoing surgical resection for MCC. On analyzing the patients in the Survival, Epidemiology, and End Results (SEER) database, we excluded all patients who either presented with distant disease, had unknown stage, had no cancer-directed therapy, or did not have histologic confirmation of MCC. This yielded a total of 1,187 patients. Since we only focused on patients with histologically confirmed tumors, all of the cases we analyzed further were in the 1988 to 2002 SEER data set. In addition, when performing the multivariate analyses we excluded all patients with incomplete/unknown information for the variables that were clinically important (n = 603). Also, Houseman et al suggested that using the SEER historic staging system is not as robust as the American Joint Committee on Cancer (AJCC). This may be true for more complex AJCC staging, but MCC AJCC staging is very similar to the SEER historic staging system. The AJCC classification for MCC designates stage I as tumors smaller than 2 cm, stage II as tumors larger than 2 cm, stage III when regional lymphatics are involved, and distant disease as stage IV. The SEER reports MCC as local, regional, or distant disease at presentation. Houseman et al suggest that including all patients rather than focusing on those alive at 4 months after diagnosis (4-month conditional survival) biases survival in favor of the radiation group. Obviously, a patient who dies before 4 months would be unlikely to have received adjuvant radiation treatment. While it is important to examine 4-month conditional survival when analyzing SEER data for most visceral malignancies for which postoperative morbidity and mortality are high, surgical morbidity and early mortality for MCC are far less likely. In addition, in our initial analyses we excluded patients with distant metastatic disease at presentation. Patients with distant disease at presentation represent a group most likely to die within 4 months of initiation of treatment. In the cohort of patients we studied, there were 1,187 patients with locoregional disease from 1988 to 2002 and only 63 died (5%) within 4 months of surgery. Housman et al performed multivariate analyses focusing on small subsets of patients and further restricted the size of the data set by including only those patients alive at 4 months after diagnosis (n = 478). The resulting subgroup analyses included far fewer patients than in the analyses included in our publication.1 As a result of the smaller number of patients, the statistical association between radiation and survival was lost. This is an example of how overinterpretation of subgroup results can be treacherous. Nonetheless, the subgroup analyses performed by Housman et al raise concerns about the validity of our published results. Reanalysis of the subgroup of patients alive at 4 months after presentation, without invoking a type II error related to small sample size, requires the use of a larger data set. Therefore, we examined the effect of radiation use on 4-month conditional survival using an expanded SEER data set, which was recently updated to include patients treated up to 2004,2 instead of the data set we used in our published results,1 which only included patients treated up to 2002. This larger data set includes 1,124 patients presenting between from 1988 to 2004. Eliminating patients with missing data yielded 962 patients. In addition to including a larger number of patients, the updated data set includes updated and longer follow-up. Cox proportional hazard multivariate analysis of the updated data set reveals an association between radiation use and 4-month conditional overall survival for patients with local and locoregional disease (P = .0173; hazard ratio, 0.889; 95% CI, 0.76 to 0.98). This finding is particularly prominent for patients with tumors larger than 2 cm in size (P = .0177). We did not find an association between the extent of surgical resection or lymphadenectomy and survival. We believe this is consistent with the findings of most other investigators. There is little data in the literature to support any association between the extent of surgical resection or lymphadenectomy and survival for MCC. In fact, Allen et al,3 which is cited by Houseman et al, does not include any analysis of the role of lymphadenectomy and survival for MCC. Despite the availability of large numbers of patients, SEER analyses are all retrospective and subject to the vagaries of selection bias. Nonetheless, the finding of a significant survival advantage associated with the use of adjuvant radiation therapy still supports the conclusion stated in the final sentence of our recent publication. "[p]rospective evaluation of adjuvant radiation in MCC is warranted." AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES
1. Mojica P, Smith D, Ellenhorn JD: Adjuvant radiation therapy is associated with improved survival in merkel cell carcinoma of the skin. J Clin Oncol 25:1043-1047, 2007 2. SEER Data, 1973-2004. http://seer.cancer.gov/data/ 3. Allen PJ, Bowne WB, Jaques DP, et al: Merkel cell carcinoma: Prognosis and treatment of patients from a single institution. J Clin Oncol 23:2300-2309, 2005
Related Correspondence
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
