Prospective evaluation of prognostic variables from patient-completed questionnaires. North Central Cancer Treatment Group

Geriatric evaluation methods in oncology and their use in clinical studies: A systematic literature review

INTRODUCTION

Therapeutic options in oncology keep on expanding. Nonetheless, older adults are underrepresented in clinical trials and those enrolled often have a better health status than their average peers, resulting in a lack of representative evidence for this heterogenous population. The inclusion of older patients and a uniform categorization of "frailty" is becoming increasingly urgent. Standardized tools could contribute to the quality and comparability of clinical trials and facilitate clinical decisions. The aim of this literature review was to elaborate an overview of the use of geriatric evaluation (GE) methods in clinical cancer research.

MATERIALS AND METHODS

We performed a literature review of the PubMed database. Clinical pharmacotherapy studies that applied or evaluated a clearly defined system for the GE of oncological patients were included. Data retrieved encompassed the applied GE method(s), cancer type(s), and pharmacotherapy investigated, the number of included patients, study type, year of publication, as well as the primary purpose of the GE. The GEs used most frequently were depicted in more depth.

RESULTS

In this literature review, 103 publications were selected for inclusion. The biggest proportion of studies (36%, n = 34) used clearly defined, but not previously validated, GE methods (study-specific GE). Standardized GE methods encountered in at least five publications were the G8 screening test (applied in 18% of included studies, n = 17), the Balducci score (7%, n = 7), and a geriatric assessment based on Hurria (5%, n = 5). The primary purpose of GE was predominantly an appraisal of its potential role in pharmacotherapy optimization. The GE also served as baseline and outcome measure, inclusion/exclusion criterion, factor for stratified randomization, and to determine treatment allocation.

DISCUSSION

The wide range of GE methods used across studies make direct comparisons difficult, and many methods are poorly characterized and/or not previously validated. The further inclusion of representative older patients in clinical trials combined with the use of a standardized GE could help clinicians in the decision-making process.

Time-dependent receiver operating characteristic curve estimator for correlated right-censored time-to-event data

In clinical trials, evaluating the accuracy of risk scores (markers) derived from prognostic models for prediction of survival outcomes is of major concern. The time-dependent receiver operating characteristic curve and the corresponding area under the receiver operating characteristic curve are appealing measures to evaluate the predictive accuracy. Several estimation methods have been proposed in the context of classical right-censored data which assumes the event time of individuals are independent. In many applications, however, this may not hold true if, for example, individuals belong to clusters or experience recurrent events. Estimates may be biased if this correlated nature is not taken into account. This paper is then aimed to fill this knowledge gap to introduce a time-dependent receiver operating characteristic curve and the corresponding area under the receiver operating characteristic curve estimation method for right-censored data that take the correlated nature into account. In the proposed method, the unknown status of censored subjects is imputed using conditional survival functions given the marker and frailty of the subjects. An extensive simulation study is conducted to evaluate and demonstrate the finite sample performance of the proposed method. Finally, the proposed method is illustrated using two real-world examples of lung cancer and kidney disease.

Privacy-preserving analysis of time-to-event data under nested case-control sampling

Analyses of distributed data networks of rare diseases are constrained by legitimate privacy and ethical concerns. Analytical centers (e.g. research institutions) are thus confronted with the challenging task of obtaining data from recruiting sites that are often unable or unwilling to share personal records of participants. For time-to-event data, recently popularized disclosure techniques with privacy guarantees (e.g., Differentially Private Generative Adversarial Networks) are generally computationally expensive or inaccessible to applied researchers. To perform the widely used Cox proportional hazards regression, we propose an easy-to-implement privacy-preserving data analysis technique by pooling (i.e. aggregating) individual records of covariates at recruiting sites under the nested case-control sampling framework before sharing the pooled nested case-control subcohort. We show that the pooled hazard ratio estimators, under the pooled nested case-control subsamples from the contributing sites, are maximum likelihood estimators and provide consistent estimates of the individual level full cohort HRs. Furthermore, a sampling technique for generating pseudo-event times for individual subjects that constitute the pooled nested case-control subsamples is proposed. Our method is demonstrated using extensive simulations and analysis of the National Lung Screening Trial data. The utility of our proposed approach is compared to the gold standard (full cohort) and synthetic data generated using classification and regression trees. The proposed pooling technique performs to near-optimal levels comparable to full cohort analysis or synthetic data; the efficiency improves in rare event settings when more controls are matched on during nested case-control subcohort sampling.

Potential of ChatGPT in facilitating research in radiation oncology?

PURPOSE

To evaluate the potential of the artificial intelligence (AI) chatbot ChatGPT in supporting young clinical scientists with scientific tasks in radio oncological research.

MATERIALS AND METHODS

Seven scientific tasks were to be completed in 3 h by 8 radiation oncologists with different scientific experience working at a university hospital: creation of a scientific synopsis, creation of a research question and corresponding clinical trial hypotheses, writing of the first paragraph of a manuscript introduction, clinical trial sample size calculation, and clinical data analyses (multivariate analysis, boxplot and survival curve). No participant had prior experience with an AI chatbot. All participants were instructed in ChatGPT v3.5 and its use was provided for all tasks. Answers were scored independently by two blinded experts. The subjective value of ChatGPT was rated by each participant. Data were analyzed with regression-, t-test and Spearman correlation (p < 0.05).

RESULTS

Participants completed tasks 1-3 with an average score of 50% and 4-7 with 56%. Scientific experience, number of original publications and of first/last authorships showed a positive correlation with overall scoring (p = 0.01-0.04). Participants with little to moderate scientific experience scored ChatGPT to be more helpful in solving tasks 4-7 compared to more experienced participants (p = 0.04), with simultaneously presenting lower scorings (p = 0.03).

CONCLUSIONS

ChatGPT did not compensate for differences in scientific experience of young clinical scientists, with less experienced researchers believing false AI-generated scientific results.

Methodologic Issues Specific to Prediction Model Development and Evaluation

Developing and evaluating statistical prediction models is challenging, and many pitfalls can arise. This article identifies what the authors believe are some common methodologic concerns that may be encountered. We describe each problem and make suggestions regarding how to address them. The hope is that this article will result in higher-quality publications of statistical prediction models.

In silico cancer immunotherapy trials uncover the consequences of therapy-specific response patterns for clinical trial design and outcome

Late-stage cancer immunotherapy trials often lead to unusual survival curve shapes, like delayed curve separation or a plateauing curve in the treatment arm. It is critical for trial success to anticipate such effects in advance and adjust the design accordingly. Here, we use in silico cancer immunotherapy trials - simulated trials based on three different mathematical models - to assemble virtual patient cohorts undergoing late-stage immunotherapy, chemotherapy, or combination therapies. We find that all three simulation models predict the distinctive survival curve shapes commonly associated with immunotherapies. Considering four aspects of clinical trial design - sample size, endpoint, randomization rate, and interim analyses - we demonstrate how, by simulating various possible scenarios, the robustness of trial design choices can be scrutinized, and possible pitfalls can be identified in advance. We provide readily usable, web-based implementations of our three trial simulation models to facilitate their use by biomedical researchers, doctors, and trialists.

'ADVANCE' (a pilot trial) ADjuVANt chemotherapy in the elderly: Developing and evaluating lower-toxicity chemotherapy options for older patients with breast cancer

INTRODUCTION

Older adults with breast cancer receiving neo/adjuvant chemotherapy are at high risk for poor outcomes and are underrepresented in clinical trials. The ADVANCE (ADjuVANt Chemotherapy in the Elderly) trial evaluated the feasibility of two neo/adjuvant chemotherapy regimens in parallel-enrolling cohorts of older patients with human epidermal growth factor receptor 2-negative breast cancer: cohort 1-triple-negative; cohort 2-hormone receptor-positive.

MATERIALS AND METHODS

Adults age ≥ 70 years with stage I-III breast cancer warranting neo/adjuvant chemotherapy were enrolled. Cohort 1 received weekly carboplatin (area under the curve 2) and weekly paclitaxel 80 mg/m2 for twelve weeks; cohort 2 received weekly paclitaxel 80 mg/m2 plus every-three-weekly cyclophosphamide 600 mg/m2 over twelve weeks. The primary study endpoint was feasibility, defined as ≥80% of patients receiving ≥80% of intended weeks/doses of therapy. All dose modifications were applied per clinician discretion.

RESULTS

Forty women (n = 20 per cohort) were enrolled from March 25, 2019 through August 3, 2020 from three centers; 45% and 35% of patients in cohorts 1 and 2 were age > 75, respectively. Neither cohort achieved targeted thresholds for feasibility. In cohort 1, eight (40.0%) met feasibility (95% confidence interval [CI] = 19.1-63.9%), while ten (50.0%) met feasibility in cohort 2 (95% CI = 27.2-72.8). Neutropenia was the most common grade 3-4 toxicity (cohort 1-65%, cohort 2-55%). In cohort 1, 80% and 85% required ≥1 dose holds of carboplatin and/or paclitaxel, respectively. In cohort 2, 10% required dose hold(s) for cyclophosphamide and/or 65% for paclitaxel.

DISCUSSION

In this pragmatic pilot examining chemotherapy regimens in older adults with breast cancer, neither regimen met target goals for feasibility. Developing efficacious and tolerable regimens for older patients with breast cancer who need chemotherapy remains an important goal.

CLINICALTRIALS

gov Identifier: NCT03858322.

Baseline Quality of Life is a Strong and Independent Prognostic Factor for Overall Survival in Metastatic Colorectal Cancer

BACKGROUND

Previous studies have established that higher baseline quality of life (QOL) scores are associated with improved survival in patients with metastatic colorectal cancer (mCRC). We examined the relationship between overall survival (OS) and baseline QOL.

PATIENTS AND METHODS

A total of 1 247 patients with mCRC participating in N9741 (comparing bolus 5-FU/LV, irinotecan [IFL] vs infusional 5-FU/leucovorin [LV]/oxaliplatin [FOLFOX] vs. irinotecan/oxaliplatin [IROX]) provided data at baseline on overall QOL using a single-item linear analogue self-assessment (LASA) 0-100 point scale. The association of OS according to clinically deficient (defined as CD-QOL, score 0-50) vs not clinically deficient (nCD-QOL, score 51-100) baseline QOL scores was tested. A multivariable analysis using Cox proportional hazards modeling was performed to adjust for the effects of multiple baseline factors. An exploratory analysis was performed evaluating OS according to baseline QOL status among patients who did or did not receive second-line therapy.

RESULTS

Baseline QOL was a strong predictor of OS for the whole cohort (CD-QOL vs nCD-QOL: 11.2 months vs 18.4 months, P < .0001), and in each arm IFL 12.4 vs 15.1 months, FOLFOX 11.1 months vs 20.6 months, and IROX 8.9 months vs 18.1 months. Baseline QOL was associated with baseline performance status (PS) (P < .0001). After adjusting for PS and treatment arm, baseline QOL was still associated with OS (P = .017).

CONCLUSIONS

Baseline QOL is an independent prognostic factor for OS in patients with mCRC. The demonstration that patient-assessed QOL and PS are independent prognostic indicators suggests that these assessments provide important complementary prognostic information.

Longitudinal Analysis of Patient-Reported Cognitive Function in Multiple Myeloma

BACKGROUND

Cancer-related cognitive impairment (CRCI) has been largely unstudied in patients with multiple myeloma (MM). This study describes patient-reported cognition over time and patient factors associated with adverse cognitive outcomes in MM.

METHODS

Participants enrolled in a registry in which they completed a geriatric assessment at study entry, and 3 & 6 months after entry. Cognitive function was assessed using the EORTC QLQ-C30 Cognitive Function subscale, with CRCI defined as scores < 75. Generalized estimating equation (GEE) models were used to fit longitudinal models to investigate differences by group and differences in changes over time by group, with adjustment for time since diagnosis.

RESULTS

One hundred and four adults with MM had mean age of 67 years and 30% identified as Black. Patient-reported CRCI was present in 18% of participants at enrollment, 21% at 3 months, and 30% at 6 months. Worse cognitive function was reported in those with impairments in physical function (P = .002), IADLs (P = .02), and performance status (P = .04), as well as in those who were prefrail/frail (P = .02) and depressed (P = .049). Greater cognitive decline over time was observed in patients without CRCI at enrollment (P < .0001) and those with lower levels of education (P = .04).

CONCLUSION

This is one of the first studies to describe longitudinal changes in patient-reported cognition in patients with MM. Several potentially intervenable factors, including physical function impairment and depression, were associated with worse cognition at study entry, but only baseline CRCI status and education level were predictive of future decline.

Prevalence and clinical correlates of cognitive impairment in adults with plasma cell disorders

INTRODUCTION

Older adults with plasma cell disorders (PCDs) experience cognitive dysfunction that may be attributable to the disease and associated therapies. Yet, this has seldom been reported in the literature. Our objectives were to describe cognitive function (objective and patient-reported) in adults with PCDs and to explore clinical correlates of cognitive impairment.

MATERIALS AND METHODS

Participants completed a geriatric assessment between March 2018 and February 2020. Cognitive function was evaluated using two objective measures - Montreal Cognitive Assessment (MoCA, cutpoint <26) and Blessed Orientation Memory Concentration Test (BOMC, cutpoint >4) - and two patient-reported outcome (PRO) measures - Patient-Reported Outcomes Measurement Information System Cognitive Function (PROMIS-CF, cutpoint <45) and European Organization for Research and Treatment of Cancer Cognitive Functioning subscale (EORTC-CF, cutpoint <75). Spearman correlations examined relationships among these measures and log binomial regression was used to examine characteristics associated with cognitive impairment, as defined by the MoCA and PROMIS-CF measures.

RESULTS

Among 86 participants with a mean age of 69 (range: 46-91), the prevalence of cognitive dysfunction was between 20% (BOMC) and 63% (MoCA). There was moderate correlation among objective measures (r = 0.51, p < 0.0001), moderate to high correlation among PRO measures (r = 0.69, p < 0.0001), but no correlation between objective and PRO measures. Factors associated with objective impairment included ≤ high school education (RR 1.46, p = 0.009), living alone (RR 1.42, p = 0.02), relapsed/refractory disease (RR 1.39, p = 0.04), empirically de-intensified induction therapy (RR 1.62, p = 0.008), frailty (RR 1.49, p = 0.04), and peripheral vascular disease (RR 1.54, p = 0.002). Factors associated with PRO impairment included social isolation (RR 3.43, p = 0.003), depression (RR 3.30, p = 0.004) and anxiety (RR 4.43, p = 0.0002), frailty (RR 3.60, p = 0.02), falls in the previous 6 months (RR 2.53, p = 0.02), and deficits in physical function (RR 4.44, p = 0.01). Older age was not associated with either objective or PRO impairment.

DISCUSSION

Cognitive impairment, using objective and PRO screening measures, was relatively common in adults with PCDs. Cancer-related factors and medical comorbidities were associated with objective cognitive impairment whereas psychosocial and functional factors were associated with PRO impairment.

Patterns and predictors of functional decline after alloHCT in older adults

BACKGROUND

As allogeneic hematopoietic cell transplantation (alloHCT) is increasingly offered to older adults, geriatric assessments (GA) have been identified as a useful tool for predicting outcomes, particularly functional status. However, very few studies have examined the longitudinal change in GA measures in the post-alloHCT period.

OBJECTIVES

The objectives of this study are to 1) describe the longitudinal change in GA and QOL measures after alloHCT and to 2) identify predictors of greater functional decline post-transplant.

STUDY DESIGN

In this single-center prospective cohort study, patients aged 50 years or older planning to undergo alloHCT completed a cancer-specific GA and the Functional Assessment of Cancer Therapy-Bone Marrow Transplant (FACT-BMT) survey at baseline prior to alloHCT and then at 3, 6, and 12 months after transplant. Changes in GA and QOL measures at each post-transplant time point (3, 6, or 12 months) compared to baseline were analyzed using paired t-tests. Exploration of potential predictors of greater post-transplant functional decline, as measured by instrumental activities of daily living (IADL) and Medical Outcomes Study Physical Health scale (MOS-PH), were examined using linear regressions and chi-squared two-sample test of proportions.

RESULTS

Mean functional status generally exhibited a pattern of decline at 3 to 6 months post-alloHCT, with recovery to near baseline by 12 months. Mean mental health and emotional QOL were lowest at baseline and improved at all time points post-transplant. Differences in baseline clinical characteristics were not associated with any differences in functional trajectories. Differences in baseline GA measures (patient-rated KPS, IADL, MOS-PH, Timed-Up-and-Go, Blessed Orientation-Memory-Concentration test, Mental Health Inventory 5) also did not predict greater functional decline at 3 months. Patients whose IADL was improved or maintained at 3 months generally maintained their functional status at 6 and 12 months. Similarly, most patients who had IADL decline at 3 months still had functional decline at 6 months, though a portion did have functional recovery by 12 months. Compared to those with improved/maintained IADL at 3 months, those with declined IADL at 3 months were significantly more likely to have persistent functional decline at 6 months (p<0.0001) and 12 months (p=0.02).

CONCLUSIONS

In older alloHCT patients, mean functional status declines short-term after alloHCT with possibility of recovery by 6 to 12 months, while mean mental and emotional health improve post-alloHCT. Functional decline at 3 months post-alloHCT is associated with persistent functional decline at 12 months.

Social support and outcomes in older adults with lung cancer

BACKGROUND

Insufficient social support is associated with increased mortality among older adults. Lung cancer is primarily a disease of older adults and is the leading cause of all cancer deaths. We assessed the association of social support with outcomes among older adults with lung cancer.

MATERIALS AND METHODS

Adults age 65 and older with lung cancer with a completed geriatric assessment (GA) were assessed. Emotional social support (ES) and tangible (material, instrumental) support (TS) measures and patient characteristics were obtained from the GA. The electronic health record was used to extract clinical variables. Simple linear regression models evaluated the association between social support scales with patient and clinical factors.

RESULTS

79 adults were assessed. White race was positively associated with ES score (p=.04), while higher BMI (p=.03), depression (p=.03) and anxiety (p=.02) were associated with worse ES. Higher BMI was associated with higher/better TS score (p=.02) while living alone was associated with lower/worse TS score (p=.03). Completion of platinum-based doublet chemotherapy with immunotherapy as planned was associated with higher ES scores (p=.02) and higher TS scores (p=.02). Disease progression was associated with lower ES scores (p=.03).

CONCLUSION

Social support may influence clinical outcomes in older adults with lung cancer. As lung cancer often portends to poor prognosis, social support may be an important prognostic indicator.

Geriatric-assessment-identified functional deficits among adults with multiple myeloma with normal performance status

OBJECTIVES

Findings from a brief geriatric assessment (GA) in a cohort of adults with multiple myeloma (MM) are presented, with particular attention to the utility of the GA in identifying important deficits in adults judged to have a normal Karnofsky Performance Status (KPS ≥ 80).

MATERIALS AND METHODS

Adults age 18 and older with MM were recruited into an observational study from 2018 to 2020. A modified Cancer and Aging Research Group (CARG) GA was administered at enrollment. Enrollees also completed the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life of Cancer Patients Core 30 questionnaire (QLQ-C30), with subscales of physical, social, role, and cognitive functioning (range 0-100; higher values indicate better function). Data were analyzed using descriptive statistics for the full cohort and stratified by concurrent KPS (score < 80 vs ≥ 80).

RESULTS

Among 89 adults, the mean age was 69.1 years, 68% were aged ≥65 years, and 70% were white. In this cohort, 78% had KPS ≥ 80. Among those with KPS ≥ 80, functional impairments (Timed Up and Go ≥14 s and dependence in ≥1 instrumental activity of daily living) were seen in 30% and 21%, respectively, with 11% reporting ≥1 fall in the prior 6 months. At least two GA-identified deficits were detected in 50% of the overall cohort and in 41% of those with KPS ≥ 80. Among those with KPS ≥ 80, self-reported physical impairment on EORTC QLQ-C30 was noted by 34%.

CONCLUSION

Using a modified CARG GA and EORTC questionnaire, functional impairments were identified among adults considered to have a good performance status based on a KPS (≥ 80). Future studies should focus on using GA measures for therapy assignment and identifying opportunities for intervening upon GA-identified deficits.

Anticipating mental health needs after chemotherapy in early-stage breast cancer using patient-reported symptom screening

PURPOSE

Many patients with breast cancer experience depression and anxiety for years after completing systemic chemotherapy, which negatively impact overall symptom burden, quality of life, and treatment outcomes. The objective of this study was to examine the utility of patient-reported outcome (PRO) measures to predict mental health needs in patients with breast cancer during post-chemotherapy follow-up care.

METHODS

In a sample of women with non-metastatic breast cancer, associations between patient-reported depression and anxiety at end of chemotherapy and post-chemotherapy mental health needs were evaluated using log-binomial regression adjusted for functional status, social activity limitations, and time from chemotherapy.

RESULTS

In a sample of 149 women, 40% reported at least mild depressive symptoms and 52% reported at least mild anxiety at the end of chemotherapy. Over an average 3.2 years post-chemotherapy (range: 0.7-5.6 years), 23% received new psychiatric diagnoses, 21% engaged in mental health specialty care, and 62% were prescribed psychotropic medications. End of chemotherapy depression and anxiety were associated with future prescription of psychotropic medications (RR 1.52; 95% CI 1.14-2.03), as well as greater number of psychotropics. Associations were strongest with serotonin-norepinephrine reuptake inhibitors [(depression: RR 4.75; 95% CI 2.06-10.95); (anxiety: RR 3.68; 95% CI 1.62-8.36); (depression and anxiety: RR 2.98; 95% CI 1.65-5.36)].

CONCLUSION

Diagnosis of and treatment for depression and anxiety are common among women with breast cancer after completing chemotherapy. Prescriptions for psychotropic medications during the initial years after systemic chemotherapy can be anticipated by depression and anxiety screening at end of chemotherapy.

K-means for shared frailty models

Background

The purpose of this research was to see how the k-means algorithm can be applied to survival analysis with single events per subject for defining groups, which can then be modeled in a shared frailty model to further allow the capturing the unmeasured confounding not already explained by the covariates in the model.

Methods

For this purpose we developed our own k-means survival grouping algorithm to handle this approach. We compared a regular shared frailty model with a regular grouping variable and a shared frailty model with a k-means grouping variable in simulations as well as analysis on a real dataset.

Results

We found that in both simulations as well as real data showed that our k-means clustering is no different than the typical frailty clustering even under different situations of varied case rates and censoring. It appeared our k-means algorithm could be a trustworthy mechanism of creating groups from data when no grouping term exists for including in a frailty term in a survival model or comparing to an existing grouping variable available in the current data to use in a frailty model.

A comparative study of forest methods for time-to-event data: variable selection and predictive performance

Background

As a hot method in machine learning field, the forests approach is an attractive alternative approach to Cox model. Random survival forests (RSF) methodology is the most popular survival forests method, whereas its drawbacks exist such as a selection bias towards covariates with many possible split points. Conditional inference forests (CIF) methodology is known to reduce the selection bias via a two-step split procedure implementing hypothesis tests as it separates the variable selection and splitting, but its computation costs too much time. Random forests with maximally selected rank statistics (MSR-RF) methodology proposed recently seems to be a great improvement on RSF and CIF.

Methods

In this paper we used simulation study and real data application to compare prediction performances and variable selection performances among three survival forests methods, including RSF, CIF and MSR-RF. To evaluate the performance of variable selection, we combined all simulations to calculate the frequency of ranking top of the variable importance measures of the correct variables, where higher frequency means better selection ability. We used Integrated Brier Score (IBS) and c-index to measure the prediction accuracy of all three methods. The smaller IBS value, the greater the prediction.

Results

Simulations show that three forests methods differ slightly in prediction performance. MSR-RF and RSF might perform better than CIF when there are only continuous or binary variables in the datasets.

For variable selection performance,

When there are multiple categorical variables in the datasets, the selection frequency of RSF seems to be lowest in most cases. MSR-RF and CIF have higher selection rates, and CIF perform well especially with the interaction term.

The fact that correlation degree of the variables has little effect on the selection frequency indicates that three forest methods can handle data with correlation.

When there are only continuous variables in the datasets, MSR-RF perform better. When there are only binary variables in the datasets, RSF and MSR-RF have more advantages than CIF.

When the variable dimension increases, MSR-RF and RSF seem to be more robustthan CIF

Conclusions

All three methods show advantages in prediction performances and variable selection performances under different situations. The recent proposed methodology MSR-RF possess practical value and is well worth popularizing. It is important to identify the appropriate method in real use according to the research aim and the nature of covariates.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01386-8.

On the stability of log-rank test under labeling errors

Motivation

Log-rank test is a widely used test that serves to assess the statistical significance of observed differences in survival, when comparing two or more groups. The log-rank test is based on several assumptions that support the validity of the calculations. It is naturally assumed, implicitly, that no errors occur in the labeling of the samples. That is, the mapping between samples and groups is perfectly correct. In this work, we investigate how test results may be affected when considering some errors in the original labeling.

Results

We introduce and define the uncertainty that arises from labeling errors in log-rank test. In order to deal with this uncertainty, we develop a novel algorithm for efficiently calculating a stability interval around the original log-rank P-value and prove its correctness. We demonstrate our algorithm on several datasets.

Availability and implementation

We provide a Python implementation, called LoRSI, for calculating the stability interval using our algorithm https://github.com/YakhiniGroup/LoRSI.

Supplementary information

Supplementary data are available at Bioinformatics online.

Study design and methods for the using exercise to relieve joint pain and improve AI adherence in older breast cancer survivors (REJOIN) trial

BACKGROUND

Aromatase Inhibitors (AIs) are recommended for survival in post-menopausal breast cancer survivors (BCS) with hormone-sensitive disease. AI Adherence is suboptimal, especially in older BCS. Joint pain is a common AI-related symptom that is associated with low AI adherence. The Using Exercise to Relieve Joint Pain in Older Breast Cancer Survivors (REJOIN) Trial will evaluate the efficacy of a self-management intervention (exercise + education) to increase knowledge/self-efficacy for symptom management, reduce joint pain and potentially increase AI adherence in older BCS planning to take AIs.

METHODS

This randomized controlled pilot trial will include sedentary BCS, 65 years and older, diagnosed with stage I-III hormone-sensitive breast cancer, who have completed primary cancer treatment and are planning to initiate AIs. We will adapt an evidence-based physical activity program for older adults that includes bi-weekly, supervised exercise sessions plus 30 min of education. The 16-week intervention program includes: 8-weeks of supervised sessions plus 8-weeks of self-guided home sessions with periodic phone coaching. We will conduct geriatric assessments plus measurements of exercise, joint pain, and AI adherence (baseline, 4, 6 and 12 months).

DISCUSSION

REJOIN is one of the first trials to exclusively target older BCS using a self-management intervention, informed by geriatric assessment and exercise physiology, to improve health outcomes in survivorship. The REJOIN trial could lay the foundation for transdisciplinary research that bridges the gap between clinical and public health perspectives in healthy aging, with the opportunity to translate clinical interventions into non-pharmacological tools for a growing, yet underserved population of older survivors.

TRIAL REGISTRATION

NCT03955627.

Optimal Supportive Care for Patients With Metastatic Breast Cancer According to Their Disease Progression Phase

The clinical progression patterns of metastatic breast cancer (MBC) are heterogeneous; patients experience acute and stable phases at different time points. The acute phase consists of rapid progressive symptomatic changes, whereas in the stable phase, patients have relatively low symptom burden. Therefore, personalized interdisciplinary care is essential. The optimal palliative or supportive care in MBC is to provide comprehensive care that is individually prioritized to the patient's disease status. The purpose of this review is to provide a practical guide for oncologists to understand the priorities for supportive care for patients with MBC in the two phases. We note that for better decision making in patient care, performance status should be broadened to consider not only physical status but also psychosocial needs and cognitive condition. We summarize the clinical importance of physical symptom control, psychosocial support, physical activity, nutrition support, and advance care planning. For optimal care, we present palliative or supportive care checklists according to the disease progression phase, combining the limited evidence with expert input. In the acute phase, close monitoring of the patient's status and symptom management take priority. In the stable phase, the focus can shift to maintenance or improvement of physical strength and emotional condition. Finally, we discuss future directions and unmet needs in providing the best supportive care for patients with MBC.

Aging Phenotypes and Restoring Functional Deficits in Older Adults With Hematologic Malignancy

BACKGROUND

Gauging fitness remains a challenge among older adults with hematologic malignancies, and interventions to restore function are lacking. We pilot a structured exercise intervention and novel biologic correlates of aging using epigenetic clocks and markers of immunosenescence to evaluate changes in function and clinical outcomes.

METHODS

Older adults (n=30) with hematologic malignancy actively receiving treatment were screened and enrolled in a 6-month exercise intervention, the Otago Exercise Programme (OEP). The impact of the OEP on geriatric assessment metrics and health-related quality of life were captured. Clinical outcomes of overall survival and hospital utilization (inpatient length of stay and emergency department use) in relationship to geriatric deficits were analyzed.

RESULTS

Older adults (median age, 75.5 years [range, 62-83 years]) actively receiving treatment were enrolled in the OEP. Instrumental activities of daily living and physical health scores (PHS) increased significantly with the OEP intervention (median PHS: visit 1, 55 [range, 0-100]; visit 2, 70 [range, 30-100]; P<.01). Patient-reported Karnofsky performance status increased significantly, and the improvement was sustained (median [range]: visit 1, 80 [40-100]; visit 3, 90 [50-100]; P=.05). Quality of life (Patient-Reported Outcome Measurement Information System [PROMIS]) improved significantly by the end of the 6-month period (median [range]: visit 1, 32.4 [19.9-47.7]; visit 3, 36.2 [19.9-47.7]; P=.01]. Enhanced measures of gait speed and balance, using the Short Physical Performance Battery scores, were associated with a 20% decrease in risk of death (hazard ratio, 0.80; 95% CI, 0.65-0.97; P=.03) and a shorter hospital length of stay (decrease of 1.29 days; 95% CI, -2.46 to -0.13; P=.03). Peripheral blood immunosenescent markers were analyzed in relationship to clinical frailty and reports of mPhenoAge epigenetic analysis are preliminarily reported. Chronologic age had no relationship to overall survival, length of stay, or emergency department utilization.

CONCLUSIONS

The OEP was effective in improving quality of life, and geriatric tools predicted survival and hospital utilization among older adults with hematologic malignancies.

A-qCPR risk score screening model for predicting 1-year mortality associated with hospice and palliative care in the emergency department

BACKGROUND

Evaluating the need for palliative care and predicting its mortality play important roles in the emergency department.

AIM

We developed a screening model for predicting 1-year mortality.

DESIGN

A retrospective cohort study was conducted to identify risk factors associated with 1-year mortality. Our risk scores based on these significant risk factors were then developed. Its predictive validity performance was evaluated using area under receiving operating characteristic analysis and leave-one-out cross-validation.

SETTING AND PARTICIPANTS

Patients aged 15 years or older were enrolled from June 2015 to May 2016 in the emergency department.

RESULTS

We identified five independent risk factors, each of which was assigned a number of points proportional to its estimated regression coefficient: age (0.05 points per year), qSOFA ⩾ 2 (1), Cancer (4), Eastern Cooperative Oncology Group Performance Status score ⩾ 2 (2), and Do-Not-Resuscitate status (3). The sensitivity, specificity, positive predictive value, and negative predictive value of our screening tool given the cutoff larger than 3 points were 0.99 (0.98-0.99), 0.31 (0.29-0.32), 0.26 (0.24-0.27), and 0.99 (0.98-1.00), respectively. Those with screening scores larger than 9 points corresponding to 64.0% (60.0-67.9%) of 1-year mortality were prioritized for consultation and communication. The area under the receiving operating characteristic curves for the point system was 0.84 (0.83-0.85) for the cross-validation model.

CONCLUSIONS

A-qCPR risk scores provide a good screening tool for assessing patient prognosis. Routine screening for end-of-life using this tool plays an important role in early and efficient physician-patient communications regarding hospice and palliative needs in the emergency department.

Serial comprehensive geriatric and quality of life assessments in adults age ≥ 50 years undergoing autologous hematopoietic cell transplantation

OBJECTIVES

We sought to examine the natural history of geriatric assessment (GA) and quality of life (QOL) domains among adults age ≥ 50 years undergoing autologous hematopoietic cell transplantation (autoHCT).

MATERIALS AND METHODS

A QOL tool and cancer-specific GA were completed before autoHCT in patients ≥50 years, and at 100 days, six months, and one year post-transplant.

RESULTS

One hundred eighty-four patients completed the pre-transplant QOL/GA assessment, 169 (92%) completed the 100-day assessment, 162 (88%) completed the six-month assessment, and 145 (79%) completed the twelve-month assessment. Functional status, as measured by instrumental activities of daily living (IADL), decreased from baseline to day 101 (mean change -0.42 points, 95% CI, -0.75 to -0.09, p = 0.01) but returned to baseline by one year. Physical function as measured by Medical Outcomes Study-Physical Health (MOS-PH) increased by mean of 3.27 points (95% CI, -0.02 to 6.56, p = 0.05) by one year. Physician-rated KPS improved by one year, but patient-rated KPS did not. No QOL metric deteriorated from baseline. Baseline factors predictive of IADL and MOS-PH as measured over time included comorbidities and disease status at transplant. IADL and MOS-PH as measured over time were not significantly associated with age.

CONCLUSIONS

AutoHCT for adults age ≥ 50 years resulted in an initial decrease in functional status, with subsequent improvement back to baseline by one year. Physical health and QOL measures were improved or unchanged over time. AutoHCT is well tolerated in well selected older patients, using patient reported geriatric metrics as outcomes.

Serial Assessment of Depression and Anxiety by Patients and Providers in Women Receiving Chemotherapy for Early Breast Cancer

BACKGROUND

Depression and anxiety are common in patients with breast cancer and associated with worse quality of life and treatment outcomes. Yet, these symptoms are often underrecognized and undermanaged in oncology practice. The objective of this study was to describe depression and anxiety severity and associated patient factors during adjuvant or neoadjuvant chemotherapy in women with early breast cancer using repeated single-item reports.

MATERIALS AND METHODS

Depression and anxiety were measured from consecutive patients and their clinicians during chemotherapy infusion visits. Associations between psychiatric symptoms and patient characteristics were assessed using Fisher's exact tests for categorical variables and t tests for continuous variables. The joint relationship of covariates significant in unadjusted analyses was evaluated using log-binomial regression. Cohen's kappa was used to assess agreement between patient- and clinician-reported symptoms.

RESULTS

In a sample of 256 patients, 26% reported at least moderately severe depression, and 41% reported at least moderately severe anxiety during chemotherapy, representing a near doubling in the prevalence of these symptoms compared with before chemotherapy. Patient-provider agreement was fair (depression: κ = 0.31; anxiety: κ = 0.28). More severe psychiatric symptoms were associated with being unmarried, having worse function, endorsing social activity limitations, using psychotropic medications, and having a mental health provider. In multivariable analysis, social activity limitations were associated with more severe depression (relative risk [RR], 2.17; 95% confidence interval [CI], 1.36-3.45) and anxiety (RR, 1.48; 95% CI, 1.05-2.09).

CONCLUSION

Oncologists frequently underestimate patients' depression and anxiety and should consider incorporating patient-reported outcomes to enhance monitoring of mental health symptoms.

IMPLICATIONS FOR PRACTICE

In this sample of 256 patients with breast cancer, depression and anxiety, measured using single-item toxicity reports completed by patients and providers, were very common during adjuvant or neoadjuvant chemotherapy. Patient-reported depression and anxiety of at least moderate severity were associated with multiple objective indicators of psychiatric need. Unfortunately, providers underrecognized the severity of their patients' mental health symptoms. The use of patient-reported, single-item toxicity reports can be incorporated into routine oncology practice and provide clinically meaningful information regarding patients' psychological health.

A robust algorithm for explaining unreliable machine learning survival models using the Kolmogorov-Smirnov bounds

A new robust algorithm based on the explanation method SurvLIME called SurvLIME-KS is proposed for explaining machine learning survival models. The algorithm is developed to ensure robustness to cases of a small amount of training data or outliers of survival data. The first idea behind SurvLIME-KS is to apply the Cox proportional hazards model to approximate the black-box survival model at the local area around a test example due to the linear relationship of covariates in the model. The second idea is to incorporate the well-known Kolmogorov-Smirnov bounds for constructing sets of predicted cumulative hazard functions. As a result, the robust maximin strategy is used, which aims to minimize the average distance between cumulative hazard functions of the explained black-box model and of the approximating Cox model, and to maximize the distance over all cumulative hazard functions in the interval produced by the Kolmogorov-Smirnov bounds. The maximin optimization problem is reduced to the quadratic program. Various numerical experiments with synthetic and real datasets demonstrate the SurvLIME-KS efficiency.

Prediction model for short-term mortality after palliative radiotherapy for patients having advanced cancer: a cohort study from routine electronic medical data

We developed a predictive score system for 30-day mortality after palliative radiotherapy by using predictors from routine electronic medical record. Patients with metastatic cancer receiving first course palliative radiotherapy from 1 July, 2007 to 31 December, 2017 were identified. 30-day mortality odds ratios and probabilities of the death predictive score were obtained using multivariable logistic regression model. Overall, 5,795 patients participated. Median follow-up was 39.6 months (range, 24.5-69.3) for all surviving patients. 5,290 patients died over a median 110 days, of whom 995 (17.2%) died within 30 days of radiotherapy commencement. The most important mortality predictors were primary lung cancer (odds ratio: 1.73, 95% confidence interval: 1.47-2.04) and log peripheral blood neutrophil lymphocyte ratio (odds ratio: 1.71, 95% confidence interval: 1.52-1.92). The developed predictive scoring system had 10 predictor variables and 20 points. The cross-validated area under curve was 0.81 (95% confidence interval: 0.79-0.82). The calibration suggested a reasonably good fit for the model (likelihood-ratio statistic: 2.81, P = 0.094), providing an accurate prediction for almost all 30-day mortality probabilities. The predictive scoring system accurately predicted 30-day mortality among patients with stage IV cancer. Oncologists may use this to tailor palliative therapy for patients.

The Effects of Dosage-Controlled Cannabis Capsules on Cancer-Related Cachexia and Anorexia Syndrome in Advanced Cancer Patients: Pilot Study

Background: Cancer-related cachexia and anorexia syndrome (CACS) is a common phenomenon in cancer patients. Cannabis has been suggested to stimulate appetite but research on this issue has yielded mixed results. The current study aimed to evaluate the effect of dosage-controlled cannabis capsules on CACS in advanced cancer patients. Methods: The cannabis capsules used in this study contained two fractions of oil-based compounds. The planned treatment was 2 × 10 mg per 24 hours for six months of tetrahydrocannabinol (THC) 9.5 mg and cannabidiol (CBD) 0.5 mg. If patients suffered from side effects, dosage was reduced to 5 mg × 2 per day (THC 4.75 mg, CBD 0.25 mg). Participants were weighed on every physician visit. The primary objective of the study was a weight gain of ≥10% from baseline. Results: Of 24 patients who signed the consent form, 17 started the cannabis capsules treatment, but only 11 received the capsules for more than two weeks. Three of six patients who completed the study period met the primary end-point. The remaining three patients had stable weights. In quality of life quaternaries, patients reported less appetite loss after the cannabis treatment (p=0.05). Tumor necrosis factor-α (TNF-α) levels decreased after the cannabis treatment but without statistical significance. According to patients’ self-reports, improvement in appetite and mood as well as a reduction in pain and fatigue was demonstrated. Conclusions: Despite various limitations, this preliminary study demonstrated a weight increase of ≥10% in 3/17 (17.6%) patients with doses of 5mgx1 or 5mgx2 capsules daily, without significant side effects. The results justify a larger study with dosage-controlled cannabis capsules in CACS.

Super Learner for Survival Data Prediction

Survival analysis is a widely used method to establish a connection between a time to event outcome and a set of potential covariates. Accurately predicting the time of an event of interest is of primary importance in survival analysis. Many different algorithms have been proposed for survival prediction. However, for a given prediction problem it is rarely, if ever, possible to know in advance which algorithm will perform the best. In this paper we propose two algorithms for constructing super learners in survival data prediction where the individual algorithms are based on proportional hazards. A super learner is a flexible approach to statistical learning that finds the best weighted ensemble of the individual algorithms. Finding the optimal combination of the individual algorithms through minimizing cross-validated risk controls for over-fitting of the final ensemble learner. Candidate algorithms may range from a basic Cox model to tree-based machine learning algorithms, assuming all candidate algorithms are based on the proportional hazards framework. The ensemble weights are estimated by minimizing the cross-validated negative log partial likelihood. We compare the performance of the proposed super learners with existing models through extensive simulation studies. In all simulation scenarios, the proposed super learners are either the best fit or near the best fit. The performances of the newly proposed algorithms are also demonstrated with clinical data examples.

Functional Status as Measured by Geriatric Assessment Predicts Inferior Survival in Older Allogeneic Hematopoietic Cell Transplantation Recipients

Allogeneic hematopoietic cell transplantation (alloHCT) has been increasingly offered to older adults with hematologic malignancies. However, optimal methods to determine fitness for alloHCT have yet to be defined. We evaluated the ability of a comprehensive geriatric assessment (CGA) to predict post-alloHCT outcomes in a single-center prospective cohort study of patients age 50 years and older. Outcomes included overall survival (OS), progression-free survival (PFS), and nonrelapse mortality (NRM). A total of 148 patients were included, with a median age of 62 years (range, 50 to 76 years). In multivariate regression analysis, several CGA measures of functional status were predictive of post-alloHCT outcomes, after adjusting for traditional prognostic factors. Any deficit in instrumental activities of daily living (IADL) was associated with inferior OS (hazard ratio [HR], 1.81, 95% confidence interval [CI], 1.07 to 3.08; P = .03) and PFS (HR, 1.85; 95% CI, 1.15 to 2.99; P = .01). A Medical Outcomes Study Physical Health scale (MOS-PH) score <85 was associated with inferior OS (HR, 1.96; 95% CI, 1.13 to 3.40; P = .02), PFS (HR, 1.75; 95% CI, 1.07 to 2.88; P = .03), and increased NRM (subdistribution HR, 2.57; 95% CI, 1.12 to 5.92; P = .03). MOS-PH score was also associated with the number of non-hematologic grade ≥3 adverse events within the first 100 days after alloHCT (rate ratio, 1.61; 95% CI, 1.04 to 2.49; P = .03). These findings support previous work suggesting that IADL is an important prognostic tool prior to alloHCT. MOS-PH is newly identified as an additional metric to identify older patients at higher risk of poor post-alloHCT outcomes, including toxicity and NRM.

Patient-reported and clinician-reported chemotherapy-induced peripheral neuropathy in patients with early breast cancer: Current clinical practice

BACKGROUND

In the current study, the authors investigated the incidence of moderate to severe chemotherapy-induced peripheral neuropathy (CIPN) for chemotherapy regimens commonly used in current clinical practice for the treatment of patients with early breast cancer. Patient-reported and clinician-assessed CIPN severity scores were compared, and risk factors for CIPN severity were identified.

METHODS

Patients completed a Patient-Reported Symptom Monitoring form and oncologists completed a Common Terminology Criteria for Adverse Events form. CIPN reports were collected prospectively during regularly scheduled infusion visits throughout the duration of chemotherapy.

RESULTS

The sample included 184 women with a mean age of 55 years; approximately 73% were white. The 4 chemotherapy regimens used were doxorubicin and cyclophosphamide plus paclitaxel (60 patients); docetaxel and cyclophosphamide (50 patients); docetaxel, carboplatin, and anti-human epidermal growth factor receptor 2 (HER2) (24 patients); and doxorubicin and cyclophosphamide plus paclitaxel and carboplatin (18 patients). All patients treated with doxorubicin and cyclophosphamide plus paclitaxel and doxorubicin and cyclophosphamide plus paclitaxel and carboplatin received paclitaxel; all patients treated with docetaxel and cyclophosphamide and docetaxel, carboplatin, and anti-HER2 received docetaxel. The chemotherapy dose was reduced in 52 patients (28%); in 15 patients (29%), this reduction was due to CIPN. Chemotherapy was discontinued in 26 patients (14%), 8 because of CIPN. Agreement between patient-reported and clinician-assessed CIPN severity scores was minimal (weighted Cohen kappa, P = .34). Patient-reported moderate to severe CIPN was higher for paclitaxel (50%) compared with docetaxel (17.7%) (P < .001). Pretreatment arthritis and/or rheumatism (relative risk [RR], 1.58; 95% CI, 1.06-2.35 [P = .023]) and regimens containing paclitaxel (RR, 2.88; 95% CI, 1.72-4.83 [P < .0001]) were associated with higher CIPN severity. Being married (RR, 0.57; 95% CI, 0.37-0.887 [P = .01]) was found to be associated with lower CIPN severity.

CONCLUSIONS

The discrepancy between patient-reported and clinician-assessed CIPN underscores the need for both patient and clinician perspectives regarding this common, dose-limiting, and potentially disabling side effect of chemotherapy.

Geriatric Assessment Predicts Hospitalization Frequency and Long-Term Care Use in Older Adult Cancer Survivors

PURPOSE

The association between geriatric assessment (GA)-identified impairments and long-term health care use in older cancer survivors remains unknown. Our objective was to evaluate whether a GA performed at cancer diagnosis was predictive of hospitalizations and long-term care (LTC) use in older adult cancer survivors.

METHODS

Older adults with GA performed between 3 months before through 6 months after diagnosis were included (N = 125). Patients with Medicare Parts A and B coverage and no managed care were identified. Hospitalizations and LTC use (skilled nursing or assisted living) were assessed up to 5 years postdiagnosis. GA risk measures were evaluated in separate Poisson models estimating the relative risk (RR) for hospital and LTC visits, adjusting for age and Charlson comorbidity score.

RESULTS

The mean age of patients was 74 years, and the majority were female (80%) and white (90%). Breast cancer (64%) and early-stage disease (stages 0 to III, 77%) were common. Prefrail/frail status (RR, 2.5; P < .001), instrumental activities of daily living impairment (RR, 5.47; P < .001), and limitations in climbing stairs (RR, 2.94; P < .001) were associated with increased hospitalizations. Prefrail/frail status (RR, 1.86; P < .007), instrumental activities of daily living impairment (RR, 4.58; P < .001), presence of falls (RR, 6.73; P < .001), prolonged Timed Up and Go (RR, 5.45; P < .001), and limitations in climbing stairs (RR, 1.89; P < .005) were associated with LTC use.

CONCLUSION

GA-identified impairments were associated with increased hospitalizations and LTC use among older adults with cancer. GA-focused interventions should be targeted toward high-risk patients to reduce long-term adverse health care use in this vulnerable population.

Functional status in a geriatric oncology setting: A review

BACKGROUND

Comprehensive Geriatric Assessment (CGA), is used in older patients with cancer to identify frailties, which can interfere with specialized treatment, and to help with therapeutic care. Functional Status (FS) is a domain of CGA in which Activities of Daily Living (ADL) and Instrumental Activities of Daily Living (IADL) are evaluation tools.

OBJECTIVE

Our study reviewed the data available on the most frequently used tools to assess ADL and IADL in a geriatric oncology setting and their predictive values on overall survival (OS), toxicity, treatment feasibility or decision and postoperative complications.

DESIGN

This review was based on a systematic search of the MEDLINE® database for articles published in English and French between January 1, 2010, and December 31, 2017. In the final analysis, 40 out of 4061 studies were included.

RESULTS

The most common ADL and IADL scales used are the Katz ADL (KL-ADL) in 25 studies and the Lawton IADL (IADL₈) in 22 studies. FS is predictive of OS in 11 out of 24 studies, chemotoxicity in 2 out of 7 studies, treatment feasibility in 2 out of 5 studies, treatment decisions in 2 out of 3 studies, and postoperative complications in 4 out of 6 studies.

CONCLUSION

FS is of prognostic value in a geriatric oncology setting despite heterogeneous methodology and inclusion criteria, in the studies included. Additional research is needed to explore more precisely the prognostic value of FS in overall survival, toxicity, treatment feasibility or decision and postoperative complications, in older cancer patients.

Evaluation of frequency of Clinical Symptoms and Signs within Six Months Prior to Death in Patients with Advanced Solid Cancers

This retrospective study documented the frequency of the clinical symptoms and signs that increase in advanced cancer patients as they move toward death in order to create a sum score and correlate it with survival. Of 572 adult patients who were treated in four selected hospitals and who died in 1998 and 1999, data at six, three, and one month(s) prior to death was available for 257. The results showed that the number of symptoms and certain clinical findings accelerated toward death, increasing the sum score. Younger patients obtained higher sum scores at one month prior to death than did elderly ones (p=0.014); this suggests that elderly patients die at a point where they show less worsening in their clinical condition than do younger patients. The score was independent of cancer type or gender. The results of this analysis provide data for further development of a clinical tool to predict long-term survival in palliative care settings.

Patient-Reported Toxicities During Chemotherapy Regimens in Current Clinical Practice for Early Breast Cancer

BACKGROUND

This study explores the incidence of patient-reported major toxicity-symptoms rated "moderate," "severe," or "very severe"-for chemotherapy regimens commonly used in early breast cancer.

PATIENTS AND METHODS

Female patients aged 21 years or older completed a validated Patient-Reported Symptom Monitoring instrument and rated 17 symptoms throughout adjuvant or neoadjuvant chemotherapy. Fisher's exact tests compared differences in percentages in symptom ratings, and general linear regression was used to model the incidence of patient-reported major toxicity.

RESULTS

In 152 patients, the mean age was 54 years (range, 24-77), and 112 (74%) were white; 51% received an anthracycline-based regimen. The proportion of patients rating fatigue, constipation, myalgia, diarrhea, nausea, peripheral neuropathy, and swelling of arms or legs as a major toxicity at any time during chemotherapy varied significantly among four chemotherapy regimens (p < .05). The mean (SD) number of symptoms rated major toxicities was 6.3 (3.6) for anthracycline-based and 4.4 (3.5) for non-anthracycline-based regimens (p = .001; possible range, 0-17 symptoms). Baseline higher body mass index (p = .03), patient-reported Karnofsky performance status ≤80 (p = .0003), and anthracycline-based regimens (p = .0003) were associated with greater total number of symptoms rated major toxicities (alternative model: chemotherapy duration, p < .0001). Twenty-six percent of dose reductions (26 of 40), 75% of hospitalizations (15 of 20), and 94% of treatment discontinuations (15 of 16) were in anthracycline-based regimens.

CONCLUSION

Capturing multiple toxicity outcomes throughout chemotherapy enables oncologists and patients to understand the range of side effects as they discuss treatment efficacies. Continuous symptom monitoring may aid in the timely development of interventions that minimize toxicity and improve outcomes. IMPLICATIONS FOR PRACTICE: This study investigated patient-reported toxicities for 17 symptoms recorded prospectively during adjuvant and neoadjuvant chemotherapy regimens for early breast cancer. An analysis of four commonly used chemotherapy regimens identified significant differences among regimens in both individual symptoms and total number of symptoms rated moderate, severe, or very severe. Longer chemotherapy regimens, such as anthracycline-based regimens followed by paclitaxel, had higher proportions of symptoms rated major toxicities. The inclusion of patient perspectives on multiple toxicity outcomes at the same time at multiple time points during chemotherapy has the potential for improving patient-provider communication regarding symptom management, patient satisfaction, and long-term clinical outcomes.

Asymptotic convergence in distribution of the area bounded by prevalence-weighted Kaplan-Meier curves using empirical process modelling

The Kaplan-Meier product-limit estimator is a simple and powerful tool in time to event analysis. An extension exists for populations stratified into cohorts where a population survival curve is generated by weighted averaging of cohort-level survival curves. For making population-level comparisons using this statistic, we analyse the statistics of the area between two such weighted survival curves. We derive the large sample behaviour of this statistic based on an empirical process of product-limit estimators. This estimator was used by an interdisciplinary National Institutes of Health-Social Security Administration team in the identification of medical conditions to prioritize for adjudication in disability benefits processing.

Patient-Reported Functional Status in Outpatients With Advanced Cancer: Correlation With Physician-Reported Scores and Survival

CONTEXT

Performance status measures are increasingly completed by patients in outpatient cancer settings, but are not well validated for this use.

OBJECTIVES

We assessed performance of a patient-reported functional status measure (PRFS, based on the Eastern Cooperative Oncology Group [ECOG]), compared with the physician-completed ECOG, in terms of agreement in ratings and prediction of survival.

METHODS

Patients and physicians independently completed five-point PRFS (lay version of ECOG) and ECOG measures on first consultation at an oncology palliative care clinic. We assessed agreement between PRFS and ECOG using weighted Kappa statistics, and used linear regression to determine factors associated with the difference between PRFS and ECOG ratings. We used the Kaplan-Meier method to estimate the patients' median survival, categorized by PRFS and ECOG, and assessed predictive accuracy of these measures using the C-statistic.

RESULTS

For the 949 patients, there was moderate agreement between PRFS and ECOG (weighted Kappa 0.32; 95% CI: 0.28-0.36). On average, patients' ratings of performance status were worse by 0.31 points (95% CI: 0.25-0.37, P < 0.0001); this tendency was greater for younger patients (P = 0.002) and those with worse symptoms (P < 0.0001). Both PRFS and ECOG scores correlated well with overall survival; the C-statistic was higher for the average of PRFS and ECOG scores (0.619) than when reported individually (0.596 and 0.604, respectively).

CONCLUSION

Patients tend to rate their performance status worse than physicians, particularly if they are younger or have greater symptom burden. Prognostic ability of performance status could be improved by using the average of patients and physician scores.

Managing Cancer Pain in Older Adults

Managing cancer pain in older adults can be complex and challenging. Understanding the unique needs of older patients with cancer is important to safe and effective pain management. The goals of this review are to discuss the assessment of older adults with cancer-related pain, treatment of cancer pain, and adverse effects or potential risks from treatment that are unique to older patients. A detailed pain assessment and when possible utilizing the geriatric assessment are vital to developing a cancer pain management plan. The geriatric assessment can help clinicians uncover problems not routinely assessed in the standard oncologic evaluation. Opioid pain medications are safe and effective for older adults with cancer pain as long as these medications are closely monitored and titrated slowly. In addition to the well-known adverse effects of opioid medications, clinicians need to be aware of the unique risks in older adults, which could include delirium, polypharmacy, and falls.

Second-generation p-values: Improved rigor, reproducibility, & transparency in statistical analyses

Verifying that a statistically significant result is scientifically meaningful is not only good scientific practice, it is a natural way to control the Type I error rate. Here we introduce a novel extension of the p-value-a second-generation p-value (pδ)-that formally accounts for scientific relevance and leverages this natural Type I Error control. The approach relies on a pre-specified interval null hypothesis that represents the collection of effect sizes that are scientifically uninteresting or are practically null. The second-generation p-value is the proportion of data-supported hypotheses that are also null hypotheses. As such, second-generation p-values indicate when the data are compatible with null hypotheses (pδ = 1), or with alternative hypotheses (pδ = 0), or when the data are inconclusive (0 < pδ < 1). Moreover, second-generation p-values provide a proper scientific adjustment for multiple comparisons and reduce false discovery rates. This is an advance for environments rich in data, where traditional p-value adjustments are needlessly punitive. Second-generation p-values promote transparency, rigor and reproducibility of scientific results by a priori specifying which candidate hypotheses are practically meaningful and by providing a more reliable statistical summary of when the data are compatible with alternative or null hypotheses.