Rematching on-the-fly: Sequential matched randomization and a case for covariate-adjusted randomization.

Covariate-adjusted randomization (CAR) can reduce the risk of covariate imbalance and, when accounted for in analysis, increase the power of a trial. Despite CAR advances, stratified randomization remains the most common CAR method. Matched randomization (MR) randomizes treatment assignment within optimally identified matched pairs based on covariates and a distance matrix. When participants enroll sequentially, sequentially matched randomization (SMR) randomizes within matches found "on-the-fly" to meet a pre-specified matching threshold. However, pre-specifying the ideal threshold can be challenging and SMR yields less-optimal matches than MR. We extend SMR to allow multiple participants to be randomized simultaneously, to use a dynamic threshold, and to allow matches to break and rematch if a better match later enrolls (sequential rematched randomization; SRR). In simplified settings and a real-world application, we assess whether these extensions improve covariate balance, estimator/study efficiency, and optimality of matches. We investigate whether adjusting for more covariates can be detrimental upon covariate balance and efficiency as is the case of traditional stratified randomization. As secondary objectives, we use the case study to assess how SMR schemes compare side-by-side with common and related CAR schemes and whether adjusting for covariates in the design can be as powerful as adjusting for covariates in a parametric model. We find each SMR extension, individually and collectively, to improve covariate balance, estimator efficiency, study power, and quality of matches. We provide a case-study where CAR schemes with randomization-based inference can be as and more powerful than non-CAR schemes with parametric adjustment for covariates.

Secondary malignancies in non-Hodgkin lymphoma survivors: 40 years of follow-up assessed by treatment modality.

BACKGROUND: Survivors of non-Hodgkin lymphoma (NHL) have increased secondary malignancy (SM) risk. We quantified this risk by patient and treatment factors. METHODS: Standardized incidence ratios (SIR, observed-to-expected [O/E] ratio) were assessed in 142,637 NHL patients diagnosed from 1975 to 2016 in the National Cancer Institute's Surveillance, Epidemiology, and End Results Program. Comparisons were made between subgroups in terms of their SIRs relative to respective endemic populations. RESULTS: In total, 15,979 patients developed SM, more than the endemic rate (O/E 1.29; p < 0.05). Compared with white patients, relative to respective endemic populations, ethnic minorities had a higher risk of SM (white O/E 1.27, 95% CI 1.25-1.29; black O/E 1.40, 95% CI 1.31-1.48; other O/E 1.59, 95% CI 1.49-1.70). Relative to respective endemic populations, patients who received radiotherapy had similar SM rates to those who did not (O/E 1.29 each), but irradiated patients had increased breast cancer (p < 0.05). Patients who received chemotherapy had higher SM rates than those who did not (O/E 1.33 vs. 1.24, p < 0.05) including more leukemia, Kaposi sarcoma, kidney, pancreas, rectal, head and neck, and colon cancers (p < 0.05). CONCLUSIONS: This is the largest study to examine SM risk in NHL patients with the longest follow-up. Treatment with radiotherapy did not increase overall SM risk, while chemotherapy was associated with a higher overall risk. However, certain subsites were associated with a higher risk of SM, and they varied by treatment, age group, race and time since treatment. These findings are helpful for informing screening and long-term follow-up in NHL survivors.

Second Primary Malignancies in Diffuse Large B-cell Lymphoma Survivors with 40 Years of Follow Up: Influence of Chemotherapy and Radiation Therapy.

Purpose: Previous studies have shown an increased risk of second primary malignancies (SPMs) in survivors of diffuse large B-cell lymphoma (DLBCL). Survivors live longer due to the intensification of and improvements in therapy; thus, we aimed to characterize SPM patterns in patients with DLBCL by treatment modality. Methods and Materials: Standardized incidence ratio and absolute excess risk of SPMs were assessed in patients with primary DLBCL from 1975 to 2016 in the National Cancer Institute's Surveillance, Epidemiology, and End Results Program. A subgroup analyses based on, sex, race, age at the time of diagnosis, latency, and treatment modality were performed. Propensity score-adjusted cumulative incidence curves were generated, stratified by treatment and accounting for death as a competing risk. Results: In total, 45,946 patients with DLBCL were identified with a mean follow up of 70 months. Overall, 9.2% of patients developed an SPM with a standardized incidence ratio of 1.23 (95% confidence interval, 1.20-1.27). There was no difference in SPM risk between men and women or Black and White patients. Patients age <25 years were particularly susceptible to the development of SPMs, with a risk 2.5 times greater than patients aged 50 to 74 years. Temporal patterns showed increasing risk of solid malignancies and decreasing risk of hematologic malignancies over time, with bladder cancer posing the greatest absolute excess risk of any cancer type after 15 years. Patients treated with radiation therapy (RT), chemotherapy (CT), and chemoradiation therapy (CRT) all had an increased risk of SPM development compared with the general population. The cumulative incidence of SPMs was the lowest in patients treated with RT and the highest when treated with CRT. In the modern treatment era, the cumulative incidence of SPM for patients treated with CT versus CRT was not significantly different. Conclusions: In this large population-based study, we demonstrate unique SPM risk patterns based on age, latency, and treatment modality that have important implications for the treatment and screening of patients diagnosed with DLBCL.

A roadmap to using randomization in clinical trials.

BACKGROUND: Randomization is the foundation of any clinical trial involving treatment comparison. It helps mitigate selection bias, promotes similarity of treatment groups with respect to important known and unknown confounders, and contributes to the validity of statistical tests. Various restricted randomization procedures with different probabilistic structures and different statistical properties are available. The goal of this paper is to present a systematic roadmap for the choice and application of a restricted randomization procedure in a clinical trial. METHODS: We survey available restricted randomization procedures for sequential allocation of subjects in a randomized, comparative, parallel group clinical trial with equal (1:1) allocation. We explore statistical properties of these procedures, including balance/randomness tradeoff, type I error rate and power. We perform head-to-head comparisons of different procedures through simulation under various experimental scenarios, including cases when common model assumptions are violated. We also provide some real-life clinical trial examples to illustrate the thinking process for selecting a randomization procedure for implementation in practice. RESULTS: Restricted randomization procedures targeting 1:1 allocation vary in the degree of balance/randomness they induce, and more importantly, they vary in terms of validity and efficiency of statistical inference when common model assumptions are violated (e.g. when outcomes are affected by a linear time trend; measurement error distribution is misspecified; or selection bias is introduced in the experiment). Some procedures are more robust than others. Covariate-adjusted analysis may be essential to ensure validity of the results. Special considerations are required when selecting a randomization procedure for a clinical trial with very small sample size. CONCLUSIONS: The choice of randomization design, data analytic technique (parametric or nonparametric), and analysis strategy (randomization-based or population model-based) are all very important considerations. Randomization-based tests are robust and valid alternatives to likelihood-based tests and should be considered more frequently by clinical investigators.

Comparative effectiveness study of patient-reported outcomes after proton therapy or intensity-modulated radiotherapy for prostate cancer.

BACKGROUND: Data continue to emerge on the relative merits of different treatment modalities for prostate cancer. The objective of this study was to compare patient-reported quality-of-life (QOL) outcomes after proton therapy (PT) and intensity-modulated radiation therapy (IMRT) for prostate cancer. METHODS: A comparison was performed of prospectively collected QOL data using the Expanded Prostate Cancer Index Composite (EPIC) questionnaire. QOL data were collected during the first 2 years after treatment for men who received PT and IMRT. PT was delivered to 1243 men at a single center at doses from 76 grays (Gy) to 82 Gy. IMRT was delivered to 204 men who were included in the Prostate Cancer Outcomes and Satisfaction with Treatment Quality Assessment (PROSTQA) study in doses from 75.6 Gy to 79.4 Gy. The Wilcoxon rank-sum test was used to compare EPIC outcomes by modality using baseline-adjusted scores at different time points. Individual questions were assessed by converting to binary outcomes and testing with generalized estimating equations. RESULTS: No differences were observed in summary score changes for bowel, urinary incontinence, urinary irritative/obstructive, and sexual domains between the 2 cohorts. However, more men who received IMRT reported moderate/big problems with rectal urgency (P = 0.02) and frequent bowel movements (P = 0.05) than men who received PT. CONCLUSIONS: There were no differences in QOL summary scores between the IMRT and PT cohorts during early follow-up (up to 2-years). Response to individual questions suggests possible differences in specific bowel symptoms between the 2 cohorts. These outcomes highlight the need for further comparative studies of PT and IMRT.

Measuring and predicting prostate cancer related quality of life changes using EPIC for clinical practice.

PURPOSE: We expanded the clinical usefulness of EPIC-CP (Expanded Prostate Cancer Index Composite for Clinical Practice) by evaluating its responsiveness to health related quality of life changes, defining the minimally important differences for an individual patient change in each domain and applying it to a sexual outcome prediction model. MATERIALS AND METHODS: In 1,201 subjects from a previously described multicenter longitudinal cohort we modeled the EPIC-CP domain scores of each treatment group before treatment, and at short-term and long-term followup. We considered a posttreatment domain score change from pretreatment of 0.5 SD or greater clinically significant and p ≤ 0.01 statistically significant. We determined the domain minimally important differences using the pooled 0.5 SD of the 2, 6, 12 and 24-month posttreatment changes from pretreatment values. We then recalibrated an EPIC-CP based nomogram model predicting 2-year post-prostatectomy functional erection from that developed using EPIC-26. RESULTS: For each health related quality of life domain EPIC-CP was sensitive to similar posttreatment health related quality of life changes with time, as was observed using EPIC-26. The EPIC-CP minimally important differences in changes in the urinary incontinence, urinary irritation/obstruction, bowel, sexual and vitality/hormonal domains were 1.0, 1.3, 1.2, 1.6 and 1.0, respectively. The EPIC-CP based sexual prediction model performed well (AUC 0.76). It showed robust agreement with its EPIC-26 based counterpart with 10% or less predicted probability differences between models in 95% of individuals and a mean ± SD difference of 0.0 ± 0.05 across all individuals. CONCLUSIONS: EPIC-CP is responsive to health related quality of life changes during convalescence and it can be used to predict 2-year post-prostatectomy sexual outcomes. It can facilitate shared medical decision making and patient centered care.

Expanded prostate cancer index composite for clinical practice: development and validation of a practical health related quality of life instrument for use in the routine clinical care of patients with prostate cancer.

PURPOSE: Measuring the health related quality of life of patients with prostate cancer in routine clinical practice is hindered by the lack of instruments enabling efficient, real-time, point of care scoring of multiple health related quality of life domains. Thus, we developed an instrument for this purpose. MATERIALS AND METHODS: The Expanded Prostate Cancer Index Composite for Clinical Practice is a 1-page, 16-item questionnaire that we constructed to measure urinary incontinence, urinary irritation, and the bowel, sexual and hormonal health related quality of life domains. We eliminated conceptually overlapping items from the 3-page Expanded Prostate Cancer Index Composite-26 and revised the questionnaire format to mirror the AUA symptom index, thereby enabling practitioners to calculate health related quality of life scores at the point of care. We administered the Expanded Prostate Cancer Index Composite for Clinical Practice to a new cohort of patients with prostate cancer in community based and academic oncology, radiation, and urology practices to evaluate instrument validity as well as ease of use in clinical practice. RESULTS: A total of 175 treated and 132 untreated subjects with prostate cancer completed the Expanded Prostate Cancer Index Composite for Clinical Practice. The domain scores of the Expanded Prostate Cancer Index Composite for Clinical Practice correlated highly with the respective domain scores from longer versions of the Expanded Prostate Cancer Index Composite (r≥0.93 for all domains). The Expanded Prostate Cancer Index Composite for Clinical Practice showed high internal consistency (Cronbach's α 0.64-0.84) and sensitivity to prostate cancer treatment related effects (p<0.05 in each of 5 health related quality of life domains). Patients completed the Expanded Prostate Cancer Index Composite for Clinical Practice efficiently (96% in less than 10 minutes and with 11% missing items). It was deemed very convenient by clinicians in 87% of routine clinical encounters and clinicians accurately scored completed questionnaires 94% of the time. CONCLUSIONS: The Expanded Prostate Cancer Index Composite for Clinical Practice is a valid instrument that enables patient reported, health related quality of life to be measured efficiently and accurately at the point of care, and thereby facilitates improved emphasis and management of patient reported outcomes.