Kaveh Zakeri

Race and competing mortality in advanced head and neck cancer

OBJECTIVES Black patients with head and neck cancer (HNC) have poorer survival and disease control compared to non-black patients, but disparities in death from non-cancer causes (i.e., competing mortality) are less well-studied. MATERIALS AND METHODS We conducted an analysis of 538 patients (169 black, 369 non-black) with stage III-IV HNC treated on one of six multi-institutional protocols between 1993 and 2004 involving multi-agent chemoradiotherapy with or without surgery. Competing mortality was defined as death due to intercurrent comorbid disease, treatment-related morbidity, or unknown cause in the absence of disease recurrence, progression, or second malignancy. Cox proportional hazards and competing risks regression were used to estimate the effect of black race on competing mortality. RESULTS Black race was associated with increased rates of comorbidity, smoking, heavy alcohol use, advanced tumor stage, and poorer performance status (p<.001 for all). Compared to non-black patients, black HNC patients had a higher 5 year cumulative incidence of disease progression (31.4%; 95% CI, 24.4-38.5% vs 23.4%; 95% CI, 19.1-28.1%) and competing mortality (28.1%; 95% CI, 21.2-35.3% vs 14.5%; 95% CI, 11.0-18.5%). When adjusting for age, male sex, body mass index, distance traveled, smoking and alcohol use, performance status, comorbidity, and tumor stage, the black race was associated with death from comorbid disease (Cox hazard ratio 2.13; 95% CI, 1.06-4.28, p=0.033). CONCLUSIONS Black patients with advanced HNC are at increased risk of both disease progression and death from competing non-cancer mortality, particularly death from comorbid disease. Improved strategies to manage comorbid disease may increase the benefit of treatment intensification in black patients.

Competing event risk stratification may improve the design and efficiency of clinical trials: Secondary analysis of SWOG 8794

BACKGROUND Composite endpoints can be problematic in the presence of competing risks when a treatment does not affect events comprising the endpoint equally. METHODS We conducted secondary analysis of SWOG 8794 trial of adjuvant radiation therapy (RT) for high-risk post-operative prostate cancer. The primary outcome was metastasis-free survival (MFS), defined as time to first occurrence of metastasis or death from any cause (competing mortality (CM)). We developed separate risk scores for time to metastasis and CM using competing risks regression. We estimated treatment effects using Cox models adjusted for risk scores and identified an enriched subgroup of 75 patients at high risk of metastasis and low risk of CM. RESULTS The mean CM risk score was significantly lower in the RT arm vs. control arm (p=0.001). The effect of RT on MFS (HR 0.70; 95% CI, 0.53-0.92; p=0.010) was attenuated when controlling for metastasis and CM risk (HR 0.76; 95% CI, 0.58-1.00; p=0.049), and the effect of RT on overall survival (HR 0.73; 95% CI, 0.55-0.96; p=0.02) was no longer significant when controlling for metastasis and CM risk (HR 0.80; 95% CI, 0.60-1.06; p=0.12). Compared to the whole sample, the enriched subgroup had the same 10-year incidence of MFS (40%; 95% CI, 22-57%), but a higher incidence of metastasis (30% (95% CI, 15-47%) vs. 20% (95% CI, 15-26%)). A randomized trial in the subgroup would have achieved 80% power with 56% less patients (313 vs. 709, respectively). CONCLUSION Stratification on competing event risk may improve the efficiency of clinical trials.

Longitudinal study of acute haematologic toxicity in cervical cancer patients treated with chemoradiotherapy.

Acute hematologic toxicity (HT) limits optimal delivery of concurrent chemoradiotherapy (CRT) for patients with pelvic malignancies. We tested the hypothesis that pelvic bone marrow (PBM) dose‐volume metrics were associated with weekly reductions in peripheral blood cell counts in cervical cancer patients undergoing CRT.

Single-Isocenter Frameless Volumetric Modulated Arc Radiosurgery for Multiple Intracranial Metastases.

BACKGROUND Stereotactic radiosurgery (SRS) is a well-accepted treatment for patients with intracranial metastases, but outcomes with volumetric modulated arc radiosurgery (VMAR) are poorly described. OBJECTIVE To report our initial clinical experience applying a novel single-isocenter technique to frameless VMAR for simultaneous treatment of multiple intracranial metastases. METHODS We performed a retrospective analysis of 15 patients undergoing frameless VMAR for multiple intracranial metastases using a single, centrally located isocenter in the period 2009 and 2011. Of these, 3 patients were treated for progressive or recurrent intracranial disease. A total of 62 metastases (median, 3 per patient; range, 2-13) were treated to a median dose of 20 Gy (range, 15-30 Gy). Three patients were treated with fractionated SRS. Follow-up including clinical examination and magnetic resonance imaging (MRI) occurred every 3 months. RESULTS The median follow-up for all patients was 7.1 months (range, 1.1-24.3), with 11 patients (73.3%) followed until death. For the remaining 4 patients alive at the time of analysis, the median follow-up was 19.6 months (range, 9.2-24.3). Local control at 6 and 12 months was 91.7% (95% confidence interval [CI], 84.6%-100.0%) and 81.5% (95% CI, 67.9%-100.0%), respectively. Regional failure was observed in 9 patients (60.0%), and 7 patients (46.7%) received salvage therapy. Overall survival at 6 months was 60.0% (95% CI, 40.3%-88.2%). Grade 3 or higher treatment-related toxicity was not observed. The median total treatment time was 7.2 minutes (range, 2.8-13.2 minutes). CONCLUSION Single-isocenter, frameless VMAR for multiple intracranial metastases is a promising technique that may provide similar clinical outcomes compared with conventional radiosurgery.

PHASE I TRIAL OF INTRAVENOUS ONCOLYTIC VACCINIA VIRUS (GL-ONC1) WITH CISPLATIN AND RADIOTHERAPY IN PATIENTS WITH LOCOREGIONALLY ADVANCED HEAD AND NECK CARCINOMA.

Purpose: Preclinical models have shown that the effectiveness of GL-ONC1, a modified oncolytic vaccinia virus, is enhanced by radiation and chemotherapy. The purpose of this study was to determine the safety of GL-ONC1 when delivered intravenously with chemoradiotherapy to patients with primary, nonmetastatic head and neck cancer. Experimental Design: Patients with locoregionally advanced unresected, nonmetastatic carcinoma of the head/neck, excluding stage III–IVA p16-positive oropharyngeal cancers, were treated with escalating doses and cycles of intravenous GL-ONC1, along with radiotherapy and chemotherapy. The primary aims were to define the MTD and dose-limiting toxicities, and to recommend a dose for phase II trials. Results: Between May 2012 and December 2014, 19 patients were enrolled. The most frequent adverse reactions included grade 1–2 rigors, fever, fatigue, and rash. Grade 3 adverse reactions included hypotension, mucositis, nausea, and vomiting. In 2 patients, the rash was confirmed as viral in origin by fluorescence imaging and viral plaque assay. In 4 patients, viral presence in tumor was confirmed on midtreatment biopsy by quantitative PCR. In 1 patient, live virus was confirmed in a tongue tumor 7 days after receiving the first dose of virus. The MTD was not reached. With median follow-up of 30 months, 1-year (2-year) progression-free survival and overall survival were 74.4% (64.1%) and 84.6% (69.2%), respectively. Conclusions: Delivery of GL-ONC1 is safe and feasible in patients with locoregionally advanced head/neck cancer undergoing standard chemoradiotherapy. A phase II study is warranted to further investigate this novel treatment strategy. Clin Cancer Res; 23(19); 5696–702. ©2017 AACR.

Improved Method to Stratify Elderly Patients With Cancer at Risk for Competing Events

PURPOSE To compare a novel generalized competing event (GCE) model versus the standard Cox proportional hazards regression model for stratifying elderly patients with cancer who are at risk for competing events. METHODS We identified 84,319 patients with nonmetastatic prostate, head and neck, and breast cancers from the SEER-Medicare database. Using demographic, tumor, and clinical characteristics, we trained risk scores on the basis of GCE versus Cox models for cancer-specific mortality and all-cause mortality. In test sets, we examined the predictive ability of the risk scores on the different causes of death, including second cancer mortality, noncancer mortality, and cause-specific mortality, using Fine-Gray regression and area under the curve. We compared how well models stratified subpopulations according to the ratio of the cumulative cause-specific hazard for cancer mortality to the cumulative hazard for overall mortality (ω) using the Akaike Information Criterion. RESULTS In each sample, increasing GCE risk scores were associated with increased cancer-specific mortality and decreased competing mortality, whereas risk scores from Cox models were associated with both increased cancer-specific mortality and competing mortality. GCE models created greater separation in the area under the curve for cancer-specific mortality versus noncancer mortality (P < .001), indicating better discriminatory ability between these events. Comparing the GCE model to Cox models of cause-specific mortality or all-cause mortality, the respective Akaike Information Criterion scores were superior (lower) in each sample: prostate cancer, 28.6 versus 35.5 versus 39.4; head and neck cancer, 21.1 versus 29.4 versus 40.2; and breast cancer, 24.6 versus 32.3 versus 50.8. CONCLUSION Compared with standard modeling approaches, GCE models improve stratification of elderly patients with cancer according to their risk of dying from cancer relative to overall mortality.

On lumping, splitting, and the nosology of clinical trial populations and end points

Published Ahead of Print on February 18, 2014 as 10.1200/JCO.2013.54.4429 The latest version is at http://jco.ascopubs.org/cgi/doi/10.1200/JCO.2013.54.4429 J OURNAL OF C LINICAL O NCOLOGY On Lumping, Splitting, and the Nosology of Clinical Trial Populations and End Points T O THE E DITOR : Wildiers et al 1 espouse conventional wisdom in reaching the conclusion that “all clinical trials in oncology should be without an upper age limit.” We agree greater represen- tation of elderly patients in clinical trials is needed but disagree with this conclusion, and feel two premises should be challenged: 1. “Clinical Trials Need to be Representative of the Whole Population in Whom the Treatment Will be Used Later” 1(p3715) Ideally, clinical trial populations should represent those in which its findings are applied; however, this is not an unqualified need. Clinical trials naturally create selective, idealized conditions that ne- cessitate extrapolating their findings to a wider population than rep- resented by their participants. 2 Furthermore, evidence that outcomes differ as a function of cancer trial participation is weak. 3 The authors emphasize that excluding elderly patients limits generalizability of treatment, and extrapolating findings from younger populations risks overutilization of expensive, ineffective treatments. However, under- utilization of effective treatments is as much a concern as overuti- lization of ineffective treatments. They consider relaxed inclusion/ exclusion criteria “the price society has to pay if it wants to ensure that older patients are not subjected to toxic therapies that provide no tangible clinical benefit.” 1(p3712) Although this seems correct, it side- steps an inconvenient truth: this can substantially increase trial costs, without greatly augmenting knowledge. Including more patients unlikely to benefit from treatment increases sample size, without necessarily increasing power. 4-8 In resource-constrained settings, op- timizing allocation strategies becomes paramount. 9 Using age as a selection factor does not imply exclusion from clinical trials generally. Different trials can be tailored to address specific patients’ needs, which might be more consistent with personalized medicine. Rather than liberalizing entry criteria, investigators should define populations most likely to benefit from treatment, to maximize efficiency and value of information gained. Reproducing trials in elderly populations assuming that results from other trials are nongeneralizable is mis- guided. Forgoing the assurance from such studies might be considered the price we pay to increase the dimensionality of scientific knowledge. 2. “Overall Survival is Considered the Gold Standard in Clinical Trials, Especially When Evaluating the Superiority of New Treatments” 1(p3712) Wildiers et al 1 do not acknowledge that overall survival (OS) is itself a composite end point, comprised of competing causes of death. They admit composite end points are justified when “the expected effects on each component are similar based on clinical/biologic plausibility.” 1(p3713 ) This condition is generally unrealistic with re- spect to OS, because typically, if a cancer therapy is to be beneficial, it Journal of Clinical Oncology, Vol 32, 2014 C O R R E S P O N D E N C E should reduce mortality from cancer, not other causes, and we hope this benefit is not offset by increased treatment-related or competing mortality. Previously, we have shown that effects on a composite end point can be interpreted as the weighted average of effects on cause- specific events. 4 As such, it is problematic when composite effects, but not cause-specific effects, are not designated explicitly, since an effect on cancer mortality or noncancer mortality tells us nothing about the effect on either outcome. Contrary to Wildiers et al’s statement that using composite end points increases statistical efficiency, their use can predispose studies to both type I 10-11 and type II error 4-8 : type I error, because a positive result may be attributable in whole or part to an effect that is inconsistent with the treatment mechanism; type II error, because a negative result may be attributable to overestimating power in the presence of competing risks. Predisposition to both type I and type II error is the calling card of a bad model, leading us to think “OS as the gold standard” 1(p3712) should go the way of the actual gold standard, formally abandoned in 1976. 12 An approach the authors briefly address, which we advocate, is to consider coprimary end points in competing risks settings. Whether type I and/or II error must be adjusted for multiple testing depends on the framework applied. For example, coprimary end points can be aggregated, assigning type I and II error for the composite end point, with cause-specific effects designated explicitly. 4,13 Wildiers et al 1 im- ply that for this approach to be valid, a precondition “is that cause of death can be reliably ascertained.” If ascertainment were biased, we agree that would be a problem, but if it were just imprecise, we are not convinced this precondition is necessary, 14 and fear it will discourage investigators from determining cause of death when it might be chal- lenging, but not impossible. A concern with composite end points is that effects on the end point may be driven in part or whole by one of its components. 15 We agree that disease-specific event probabilities should be reported, but would add that competing event probabilities, and effects on competing events should be reported, to safeguard against publication bias. Since it is not typically reasonable to hypoth- esize that a cancer therapy will favorably affect noncancer mortality, we should view claims that treatments improve survival in the elderly with utmost scrutiny, particularly when the mechanism for this ben- efit cannot be evinced using clinical data. The authors list increased sample size among the limitations of coprimary end points; in con- trast, we consider this a justifiable cost to verify the mechanism by which beneficial effects are achieved. Specificity is the hallmark of every science, and one we should aspire to in clinical oncology. In general, we believe readers should be wary of rules for defining trial populations and end points without reference to a specific question. Age limits should not be off-limits for clinical trials. If one’s hypothesis is that effects are homogeneous with respect to age, then it makes no difference if elderly patients are included, save concerns regarding efficiency. If one’s hypothesis is that effects are heterogeneous, then elderly subpopulations could be stud- ied separately. Generalizability in geriatric subpopulations is most likely to be a significant concern when there is reason to expect the cause-specific treatment effects (not composite effects) would differ

Cost-effectiveness Analysis of Nivolumab for Treatment of Platinum-Resistant Recurrent or Metastatic Squamous Cell Carcinoma of the Head and Neck

Background The CheckMate 141 trial found that nivolumab improved survival for patients with recurrent or metastatic head and neck cancer (HNC). Despite the improved survival, nivolumab is much more expensive than standard therapies. This study assesses the cost-effectiveness of nivolumab for the treatment of HNC. Methods We constructed a Markov model to simulate treatment with nivolumab or standard single-agent therapy for patients with recurrent or metastatic platinum-refractory HNC. Transition probabilities, including disease progression, survival, and probability of toxicity, were derived from clinical trial data, while costs (in 2017 US dollars) and health utilities were estimated from the literature. Incremental cost-effectiveness ratios (ICERs), expressed as dollar per quality-adjusted life-year (QALY), were calculated, with values of less than

Radiation Oncology in the 21st Century: Prospective Randomized Trials That Changed Practice… or Didn’t!

100 000/QALY considered cost-effective from a health care payer perspective. We conducted one-way and probabilistic sensitivity analyses to assess model uncertainty. Results Our base case model found that treatment with nivolumab increased overall cost by

Specificity of Genetic Biomarker Studies in Cancer Research: A Systematic Review.

117 800 and improved effectiveness by 0.400 QALYs compared with standard therapy, leading to an ICER of