Originally published as JCO Early Release 10.1200/JCO.2009.23.8519 on August 31 2009

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Google Scholar Articles by Detterbeck, F. C. PubMed PubMed Citation Articles by Detterbeck, F. C. Related Articles Related Article Social Bookmarking                            What's this?

Critical Points When Using Vectors

Section of Thoracic Surgery, Yale University School of Medicine, New Haven, CT

To the Editor:

I read with interest the article by Zhu et al¹ of using support vector machines to identify prognostic classifiers for stage IB non–small-cell lung cancer (NSCLC). This is an interesting technique, and the authors have done a lot of work. However, I am worried that the study design has two significant flaws.

First, while it is true that the 5-year survival of patients with resected stage pIB NSCLC is 60% to 70%, it is important to take into account that only about half of the deaths are due to recurrence of the lung cancer.^2–8 The other 40% to 45% of deaths are due to such things as cardiovascular diseases and new primary cancers. This means that factors predictive of death for any reason cannot be interpreted as an indication to give adjuvant chemotherapy. The poor prognostic classifiers may be indicating that the patients should have their cholesterol lowered, or have a colonoscopy, and so on. It would be much better to do an analysis looking for markers of a poor risk of lung cancer recurrence. Many, if not most, articles discussing the prognosis of stage I NSCLC make the mistake of equating the lack of survival with lung cancer death. For most of these articles, however, it is merely an erroneous or deliberately misleading statement, perhaps in order to play on the fear of recurrence. In Zhu et al,¹ however, it seems to me this mistake results in a fundamental flaw of the study, not just a misleading statement.

Second, the training and validation sets of patients are from two different time periods. Ideally, the training set should be picked randomly from the total sample, and the validation done in the remaining patients. Why was the study not done this way? Having said this, it is likely that the use of different cohorts of patients (from different time periods) would make the correlation between the training and validation cohort worse; the fact that it was good suggests that the effect of the different time periods was not so great.

Along these lines, I am confused by the contradiction between the statement that the study involved a total of 148 consecutive patients, and the later statement that it involved 73 randomly selected patients between 1990 and 1998 and 75 randomly selected patients between 1998 and 2005.¹ It doesn't make sense to randomly select all of the patients. It seems to me they were selected by time period, and not randomly. Can this be clarified?

AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

REFERENCES

1. Zhu Z-H, Sun B-Y, Ma Y, et al: Three immunomarker support vector machines–based prognostic classifiers for stage IB non–small-cell lung cancer. J Clin Oncol 27:1091–1099.

2. Thomas P, Rubinstein L; the Lung Cancer Study Group. Cancer recurrence after resection: T1 N0 non–small-cell lung cancer. Ann Thorac Surg 49:242–247, 1990.[Abstract]

3. Martini N, Bains MS, Burt ME, et al: Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 109:120–129, 1995.[Abstract/Free Full Text]

4. Read RC, Schaefer R, North N, et al: Diameter, cell type, and survival in stage I primary non–small-cell lung cancer. Arch Surg 123:446–449, 1988.[Abstract/Free Full Text]

5. Pairolero PC, Williams DE, Bergstralh EJ, et al: Postsurgical stage I bronchogenic carcinoma: Morbid implications of recurrent disease. Ann Thorac Surg 38:331–338, 1984.[Abstract]

6. Harpole DH Jr, Herndon JE II, Young WG Jr, et al: Stage I non–small-cell lung cancer: A multivariate analysis of treatment methods and patterns of recurrence. Cancer 76:787–796, 1995.[CrossRef][Medline]

7. Little AG, DeMeester TR, Ferguson MK, et al: Modified stage I (T1N0M0, T2N0M0), non–small-cell lung cancer: Treatment results, recurrence patterns, and adjuvant immunotherapy Surgery, 100:621–628, 1986.

8. Macchiarini P, Fontanini G, Hardin MJ, et al: Blood vessel invasion by tumor cells predicts recurrence in completely resected T1 N0 M0 non–small-cell lung cancer. J Thorac Cardiovasc Surg 106:80–89, 1993.[Abstract]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Article

• Reply to F.C. Detterbeck
  Tie-Hua Rong and Zhi-Hua Zhu
  JCO 2009 27: 167 [Full Text]

This article has been cited by other articles:

				T.-H. Rong and Z.-H. Zhu Reply to F.C. Detterbeck J. Clin. Oncol., October 20, 2009; 27(30): e167 - e167. [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Google Scholar Articles by Detterbeck, F. C. PubMed PubMed Citation Articles by Detterbeck, F. C. Related Articles Related Article Social Bookmarking                            What's this?