Duration Studies in Oncology Have Often Been Poorly Designed and Analyzed

Get Permission

The ASCO Post is pleased to introduce “At Microphone 1,” an occasional column written by Steven E. Vogl, MD, of Bronx, New York. When he’s not in his clinic, Dr. Vogl can generally be found at major oncology meetings, and often at the microphone where he stands ready with important questions for presenters of new data. Here Dr. Vogl shares his thoughts on duration trials in oncology, in particular one presented during the plenary session at ASCO 2017. 

Steven E. Vogl, MD

Steven E. Vogl, MD

The recent IDEA combined analysis of six trials examining the duration of adjuvant oxaliplatin-based therapy for patients with resected node-positive colon cancer1 makes clear again how difficult such studies have proved in the past. 

Many patients stop their assigned treatment early— both in shorter and longer assigned groups. A few patients keep going beyond their assigned durations. Dose reductions, delays, and interruptions bedevil interpretations. Some of these deviations from planned therapy are inevitable, but some of the major problems can be avoided by randomizing patients at the point treatments diverge. For the IDEA studies, this would have been after 3 months of leucovorin, oxaliplatin, and fluorouracil (FOLFOX) or capecitabine and oxaliplatin (CapeOx). Thus, any patients stopping therapy before the 3-month point would not have been randomized. 

Dr. Vogl is an oncologist who works in the community oncology setting. He treats most adult cancers and manages to do so entirely in the office setting. He is affiliated with Montefiore Medical Center in New York City and White Plains Hospital.

About 12% of patients assigned to 3 months of chemotherapy stopped all therapy before the treatment arms diverged (10% of FOLFOX and 14% of CapeOx). Most likely, 12% of those assigned to 6 months of therapy stopped similarly (Dr. Shi’s presentation at the 2017 ASCO Annual Meeting did not give us this information.) The 12% of early stoppers drag down the outcome results for both arms and dilute any advantage of longer therapy, because they make the two arms more alike. This could lead to a false conclusion of noninferiority. Also dragging down the potential for finding a difference was that many patients assigned to 6 months of treatment did not finish the planned therapy: 29% on FOLFOX and 35% on CapeOx. More than one-third of these patients had stopped therapy in the first 3 months. It is not too late to repeat the analysis deleting any patients who stopped chemotherapy before the completion of 3 months of treatment. 

“Classic” and recent examples of trial “death by premature randomization” are rife in the oncology literature. 

RTOG 94-05 Trial: Localized Esophageal Cancer 

THE CLASSIC example of study destruction by early entry and randomization is Radiation Therapy Oncology Group (RTOG) 94-05 (Intergroup 0123).2 This study looked at dose (and therefore duration, since the daily fraction size was fixed in this study) of external-beam radiation as part of curative chemoradiation for primary esophageal cancer. The current standard dose remains 50 cGy for this disease, because the group assigned to 60 cGy in this study had more early deaths than the control group. These deaths occurred before the patients ever reached a cumulative dose of 50 cGy (the full dose in the control arm). The study had to be closed for futility after 236 patients were enrolled (no matter what happened to patients accrued later, the longer therapy/higher-dose arm had already done so poorly the longer/higher-dose treatment had a < 2.5% chance of proving better than the standard-dose arm). In their discussion, the authors admitted the design error. They attributed the error to concern that accrual would be hindered if informed consent had to be obtained after 5 weeks of toxic chemoradiation therapy. 

Had RTOG 94-05 randomized after 50 cGy, our patients might now be benefiting from higher-dose, more-effective radiation therapy.
— Steven E. Vogl, MD

Tweet this quote

Our patients with esophageal cancer receiving chemoradiation therapy with curative intent may still be paying the price for this poorly designed study—they may be receiving a suboptimal dose of radiation. The general assumption with radiation therapy is that higher doses work better, provided they are not too toxic! A dose of 60 cGy to the mediastinum is now standard with chemotherapy for stage III non–small cell lung cancer (NSCLC)—it is not too toxic. Had RTOG 94-05 randomized after 50 cGy, our patients might now be benefiting from higher-dose, more-effective radiation therapy. 

KCSG-LU05-04 Trial: Stage III Lung Cancer 

TREATMENT OF patients with stage III NSCLC was the subject of a similar design mistake: the Korean Cancer Study Group (KCSG)-LU05-04 trial was conducted in Korea, Japan, and Taiwan and presented at the 2014 ASCO Annual Meeting by Keunchil Park, MD, PhD, of Samsung Medical Center in Seoul, South Korea.3 This study looked at adjuvant chemotherapy after chemoradiation therapy for inoperable disease. Most of us who work in the field believe the low-dose radiosensitizing chemotherapy given during radiation therapy is inadequate to sterilize micrometastases, so more chemotherapy given in full doses after chemoradiation therapy should help, as does full systemic chemotherapy for resected stages II and III disease. 

Results reported to date have shown added toxicity from added chemotherapy but no added benefit. The well-meant effort of the Korean Cancer Study Group failed to show a benefit from adjuvant docetaxel and cisplatin in overall or progression-free survival. Alas, 31.6% of patients randomized before the start of chemoradiation therapy never made it to the consolidation portion of the study. The analysis presented in 2014, however, included all of them. The diluting effect of these patients who did poorly and never made it to “consolidation” chemotherapy may have obscured a clinically significant benefit for the latter. 

Breast Cancer Clinical Trials 

NOT ALL studies of treatment duration randomized patients long before the treatment programs diverged. Classic studies of extended adjuvant aromatase inhibitor therapy for breast cancer—MA.17,4 MA.17R,5 and the recently reported NSABP B-426—all randomized trials of aromatase inhibitor vs observation at the point of divergence of the two arms. Compliance with the oral aromatase inhibitor was a big issue in all of these studies, but not prior to the divergence of the treatment arms, since randomization was only after completion of the portion of adjuvant therapy common to both arms. 

The IDEA collaborators should reanalyze their results excluding from both arms those patients who stopped therapy for any reason before the 3-month point was reached.
— Steven E. Vogl, MD

Tweet this quote

Two recently reported Dutch studies of the duration of aromatase inhibitors—DATA7 and IDEAL8— randomized patients 2.5 to 3 years before treatment arms diverged. Each had major rates of treatment cessation before the arms diverged. The IDEAL investigators reported an analysis deleting early stoppers from both arms. The DATA investigators did not—perhaps this contributed to DATA’s failure to demonstrate any advantage from longer therapy with aromatase inhibitors.9 

Where to Go From Here 

THE IDEA collaborators should reanalyze their results excluding from both arms those patients who stopped therapy for any reason before the 3-month point was reached. This will emphasize the difference between the arms and perhaps reveal greater advantages to longer adjuvant chemotherapy. 

The next generation of trials looking at therapy duration should randomize only after the common period of therapy is complete. For chemotherapy trials, stratification should include events during this common period, including dose and toxicity, as well as patient condition at randomization. ■

DISCLOSURE: Dr. Vogl reported no conflicts of interest. 


1. Shi Q, Sobrero AF, Shields AF, et al: Prospective pooled analysis of six phase III trials investigating duration of adjuvant oxaliplatin-based therapy (3 vs 6 months) for patients with stage III colon cancer: The IDEA (International Duration Evaluation of Adjuvant chemotherapy) collaboration. 2017 ASCO Annual Meeting. Abstract LBA1. Presented June 4, 2017. 

2. Minsky BD, Pajak TF, Ginsberg RJ, et al: INT 0123 (Radiation Therapy Oncology Group 94-05) phase III trial of combined-modality therapy for esophageal cancer: High-dose versus standard-dose radiation therapy. J Clin Oncol 20:1167- 1174, 2002

3. Ahn JS, Ahn YC, Kim JH, et al: Multinational randomized phase III trial with or without consolidation chemotherapy using docetaxel and cisplatin after concurrent chemoradiation in inoperable stage III non-small-cell lung cancer: KCSG-LU05-04. J Clin Oncol 33:2660-2666, 2015

4. Goss PE, Ingle JN, Martino S, et al: A randomized trial of letrozole in postmenopausal women after 5 years of tamoxifen therapy for early-stage breast cancer. N Engl J Med 349:1793-1802, 2003

5. Goss PE, Ingle JN, Pritchard KI, et al: Extending aromatase-inhibitor adjuvant therapy to 10 years. N Engl J Med 375:209-219, 2016

6. Mamounas EP, Bandos H, Lembersky BC, et al: A randomized, double-blinded, placebo-controlled clinical trial of extended adjuvant endocrine therapy with letrozole in postmenopausal women with hormone-receptor-positive breast cancer who have completed previous adjuvant treatment with an aromatase inhibitor. 2016 San Antonio Breast Cancer Symposium. Abstract S1-05

7. Tjan-Heijnen VC, Van Hellemond IE, Peer PG, et al: S1-03: First results from the multicenter phase III DATA study comparing 3 versus 6 years of anastrozole after 2-3 years of tamoxifen in postmenopausal women with hormone receptor–positive early breast cancer. 2016 San Antonio Breast Cancer Symposium. Abstract S1-03

8. Blok EJ, Van de Velde CJH, Meershoek-Klein Kranenbarg EM, et al: Optimal duration of extended letrozole treatment after 5 years of adjuvant endocrine therapy. 2016 San Antonio Breast Cancer Symposium. Abstract S1-04

9. Vogl S: Value of extended adjuvant AI therapy is still unclear. Clin Oncol News February 13, 2017. Available at http:// NY-Value-of-Extended-Adjuvant-AI-Therapy-Is-Still-Unclear/40362/ses=ogst?enl=true. Accessed September 6, 2017.