Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies.

Rapid advances in automated methods for extracting large numbers of quantitative features from medical images have led to tremendous growth of publications reporting on radiomic analyses. Translation of these research studies into clinical practice can be hindered by biases introduced during the design, analysis, or reporting of the studies. Herein, the authors review biases, sources of variability, and pitfalls that frequently arise in radiomic research, with an emphasis on study design and statistical analysis considerations. Drawing on existing work in the statistical, radiologic, and machine learning literature, approaches for avoiding these pitfalls are described.

18F-fluorodeoxyglucose (FDG) PET or 18F-fluorothymidine (FLT) PET to assess early response to aromatase inhibitors (AI) in women with ER+ operable breast cancer in a window-of-opportunity study.

PURPOSE: This study evaluated the ability of 18F-Fluorodeoxyglucose (FDG) and 18F-Fluorothymidine (FLT) imaging with positron emission tomography (PET) to measure early response to endocrine therapy from baseline to just prior to surgical resection in estrogen receptor positive (ER+) breast tumors. METHODS: In two separate studies, women with early stage ER+ breast cancer underwent either paired FDG-PET (n = 22) or FLT-PET (n = 27) scans prior to endocrine therapy and again in the pre-operative setting. Tissue samples for Ki-67 were taken for all patients both prior to treatment and at the time of surgery. RESULTS: FDG maximum standardized uptake value (SUVmax) declined in 19 of 22 lesions (mean 17% (range -45 to 28%)). FLT SUVmax declined in 24 of 27 lesions (mean 26% (range -77 to 7%)). The Ki-67 index declined in both studies, from pre-therapy (mean 23% (range 1 to 73%)) to surgery [mean 8% (range < 1 to 41%)]. Pre- and post-therapy PET measures showed strong rank-order agreement with Ki-67 percentages for both tracers; however, the percent change in FDG or FLT SUVmax did not demonstrate a strong correlation with Ki-67 index change or Ki-67 at time of surgery. CONCLUSIONS: A window-of-opportunity approach using PET imaging to assess early response of breast cancer therapy is feasible. FDG and FLT-PET imaging following a short course of neoadjuvant endocrine therapy demonstrated measurable changes in SUVmax in early stage ER+ positive breast cancers. The percentage change in FDG and FLT-PET uptake did not correlate with changes in Ki-67; post-therapy SUVmax for both tracers was significantly associated with post-therapy Ki-67, an established predictor of endocrine therapy response.

Whole-Body Characterization of Estrogen Receptor Status in Metastatic Breast Cancer with 16α-18F-Fluoro-17ß-Estradiol Positron Emission Tomography: Meta-Analysis and Recommendations for Integration into Clinical Applications.

Estrogen receptor (ER) status by immunohistochemistry (IHC) of cancer tissue is currently used to direct endocrine therapy in breast cancer. Positron emission tomography (PET) with 16α-18F-fluoro-17ß-estradiol (18 F-FES) noninvasively characterizes ER ligand-binding function of breast cancer lesions. Concordance of imaging and tissue assays should be established for 18 F-FES PET to be an alternative or complement to tissue biopsy for metastatic lesions. We conducted a meta-analysis of published results comparing 18 F-FES PET and tissue assays of ER status in patients with breast cancer. PubMed and EMBASE were searched for English-language manuscripts with at least 10 patients and low overall risk of bias. Thresholds for imaging and tissue classification could differ between studies but had to be clearly stated. We used hierarchical summary receiver-operating characteristic curve models for the meta-analysis. The primary analysis included 113 nonbreast lesions from 4 studies; an expanded analysis included 327 total lesions from 11 studies. Treating IHC results as the reference standard, sensitivity was 0.78 (95% confidence region 0.65-0.88) and specificity 0.98 (0.65-1.00) for the primary analysis of nonbreast lesions. In the expanded analysis including non-IHC tissue assays and all lesion sites, sensitivity was 0.81 (0.73-0.87) and specificity 0.86 (0.68-0.94). These results suggest that 18 F-FES PET is useful for characterization of ER status of metastatic breast cancer lesions. We also review current best practices for conducting 18 F-FES PET scans. This imaging assay has potential to improve clinically relevant outcomes for patients with (historically) ER-positive metastatic breast cancer, including those with brain metastases and/or lobular histology. IMPLICATIONS FOR PRACTICE: 16α-18F-fluoro-17ß-estradiol positron emission tomography (18 F-FES PET) imaging assesses estrogen receptor status in breast cancer in vivo. This work reviews the sensitivity and specificity of 18 F-FES PET in a meta-analysis with reference tissue assays and discusses best practices for use of the tracer as an imaging biomarker. 18 F-FES PET could enhance breast cancer diagnosis and staging as well as aid in therapy selection for patients with metastatic disease. Tissue sampling limitations, intrapatient heterogeneity, and temporal changes in molecular markers make it likely that 18 F-FES PET will complement existing assays when clinically available in the near future.

Caveolin-1 promotes the tumor suppressor properties of oncogene-induced cellular senescence.

Oncogene-induced senescence (OIS) is considered a powerful tumor suppressor mechanism. Caveolin-1 acts as a scaffolding protein to functionally regulate signaling molecules. We demonstrate that a lack of caveolin-1 expression inhibits oncogenic K-Ras (K-RasG12V)-induced premature senescence in mouse embryonic fibroblasts and normal human bronchial epithelial cells. Oncogenic K-Ras induces senescence by limiting the detoxification function of MTH1. We found that K-RasG12V promotes the interaction of caveolin-1 with MTH1, which results in inhibition of MTH1 activity. Lung cancer cells expressing oncogenic K-Ras have bypassed the senescence barrier. Interestingly, overexpression of caveolin-1 restores cellular senescence in both A549 and H460 lung cancer cells and inhibits their transformed phenotype. In support of these findings, our in vivo data demonstrate that overexpression of oncogenic K-Ras (K-RasG12D) induces cellular senescence in the lung of wildtype but not caveolin-1-null mice. A lack of K-RasG12D-induced premature senescence in caveolin-1-null mice results in the formation of more abundant lung tumors. Consistent with these data, caveolin-1-null mice overexpressing K-RasG12D display accelerated mortality. Finally, our animal data were supported by human sample analysis in which we show that caveolin-1 expression is dramatically down-regulated in lung adenocarcinomas from lung cancer patients, both at the mRNA and protein levels, and that low caveolin-1 expression is associated with poor survival. Together, our data suggest that lung cancer cells escape oncogene-induced premature senescence through down-regulation of caveolin-1 expression to progress from premalignant lesions to cancer.

Comparison of PD-L1 immunohistochemistry assays and response to PD-1/L1 inhibitors in advanced non-small-cell lung cancer in clinical practice.

AIMS: Several studies have demonstrated analytical comparability between different PD-L1 assays, but their clinical validity in non-small-cell lung cancer in terms of response to treatment outside clinical trials has not been established. The aim of our study is to assess the analytical performance of laboratory-developed tests for Ventana SP263 and Agilent/Dako 22C3, and to investigate the association between PD-L1 assays and response to PD-1/L1 inhibitors. METHODS AND RESULTS: PD-L1 SP263 and 22C3 assays were performed on 302 consecutive non-small-cell lung carcinoma samples Both assays were optimised for use on the automated Ventana BenchMark Ultra platform. Scoring algorithms for staining of the tumour cells using the established cut-offs were applied to all samples. Best overall response (BOR) for 44 patients treated with either nivolumab, pembrolizumab or atezolizumab were assessed using recist version 1.1 and correlated with PD-L1 assay results. Assays showed good agreement, with a concordance correlation coefficient of 0.86 [95% confidence interval (CI) = 0.82-0.90)]. Comparing the assays using cut-offs of 1%, 5%, 10%, 1-49% and ≥50% showed an association between the two assays (P < 0.0001). The SP263 10% cut-off (P = 0.032) was associated with BOR, whereas the 1% (P = 0.087) and 5% (P = 0.051) cut-offs were not. In contrast 22C3, cut-offs of 1% (P = 0.019), 5% (P = 0.025) and 10% (P = 0.014) were all associated with BOR. CONCLUSIONS: The SP263 and 22C3 LDTs demonstrated good analytical concordance, and correlation with response to PD-1/L1 inhibitors.

Insulin receptor substrate-1 deficiency drives a proinflammatory phenotype in KRAS mutant lung adenocarcinoma.

Insulin receptor substrate-1 (IRS-1) is a signaling adaptor protein that interfaces with many pathways activated in lung cancer. It has been assumed that IRS-1 promotes tumor growth through its ability to activate PI3K signaling downstream of the insulin-like growth factor receptor. Surprisingly, tumors with reduced IRS-1 staining in a human lung adenocarcinoma tissue microarray displayed a significant survival disadvantage, especially within the Kirsten rat sarcoma viral oncogene homolog (KRAS) mutant subgroup. Accordingly, adenoviral Cre recombinase (AdCre)-treated LSL-Kras/Irs-1(fl/fl) (Kras/Irs-1(-/-)) mice displayed increased tumor burden and mortality compared with controls. Mechanistically, IRS-1 deficiency promotes Janus kinase/signal transducers and activators of transcription (JAK/STAT) signaling via the IL-22 receptor, resulting in enhanced tumor-promoting inflammation. Treatment of Kras/Irs-1(+/+) and Kras/Irs-1(-/-) mice with JAK inhibitors significantly reduced tumor burden, most notably in the IRS-1-deficient group.

Breast MRI contrast enhancement kinetics of normal parenchyma correlate with presence of breast cancer.

BACKGROUND: We investigated dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) contrast enhancement kinetic variables quantified from normal breast parenchyma for association with presence of breast cancer, in a case-control study. METHODS: Under a Health Insurance Portability and Accountability Act compliant and Institutional Review Board-approved protocol, DCE-MRI scans of the contralateral breasts of 51 patients with cancer and 51 controls (matched by age and year of MRI) with biopsy-proven benign lesions were retrospectively analyzed. Applying fully automated computer algorithms on pre-contrast and multiple post-contrast MR sequences, two contrast enhancement kinetic variables, wash-in slope and signal enhancement ratio, were quantified from normal parenchyma of the contralateral breasts of both patients with cancer and controls. Conditional logistic regression was employed to assess association between these two measures and presence of breast cancer, with adjustment for other imaging factors including mammographic breast density and MRI background parenchymal enhancement (BPE). The area under the receiver operating characteristic curve (AUC) was used to assess the ability of the kinetic measures to distinguish patients with cancer from controls. RESULTS: When both kinetic measures were included in conditional logistic regression analysis, the odds ratio for breast cancer was 1.7 (95 % CI 1.1, 2.8; p = 0.017) for wash-in slope variance and 3.5 (95 % CI 1.2, 9.9; p = 0.019) for signal enhancement ratio volume, respectively. These odds ratios were similar on respective univariate analysis, and remained significant after adjustment for menopausal status, family history, and mammographic density. While percent BPE was associated with an odds ratio of 3.1 (95 % CI 1.2, 7.9; p = 0.018), in multivariable analysis of the three measures, percent BPE was non-significant (p = 0.897) and the two kinetics measures remained significant. For the differentiation of patients with cancer and controls, the unadjusted AUC was 0.71 using a combination of the two measures, which significantly (p = 0.005) outperformed either measure alone (AUC = 0.65 for wash-in slope variance and 0.63 for signal enhancement ratio volume). CONCLUSIONS: Kinetic measures of wash-in slope and signal enhancement ratio quantified from normal parenchyma in DCE-MRI are jointly associated with presence of breast cancer, even after adjustment for mammographic density and BPE.

Diffusion-Weighted Breast Magnetic Resonance Imaging: A Semiautomated Voxel Selection Technique Improves Interreader Reproducibility of Apparent Diffusion Coefficient Measurements.

OBJECTIVE: To determine whether a semiautomated voxel selection technique improves interreader reproducibility for breast apparent diffusion coefficient (ADC) measurements. METHODS: Three readers retrospectively performed ADC measurements of 31 breast lesions (16 malignant, 15 benign) and contralateral normal tissue in 26 women both unassisted (manual method) and assisted by a semiautomated software tool that excludes voxels below a dynamically specified signal intensity threshold. Reproducibility between readers for each technique was assessed by Bland-Altman analysis and concordance correlation coefficients (CCCs). RESULTS: Differences between readers' measured ADCs of lesions were smaller with the semiautomated tool vs the manual method. Concordance correlation coefficients for each reader pair were greater with the semiautomated tool for lesions (mean CCC difference, 0.11; 95% confidence interval, 0.04-0.26). For normal tissue, reader agreement was lower than for lesions and did not differ based on software tools (mean CCC difference, 0.00; 95% confidence interval, -0.14 to 0.13). CONCLUSIONS: A semiautomated voxel selection tool can improve interreader reproducibility of breast lesion ADC measures.

Clinicopathologic features and outcomes of patients with lung adenocarcinomas harboring BRAF mutations in the Lung Cancer Mutation Consortium.

BACKGROUND: The advent of effective targeted therapy for BRAF(V600E) -mutant lung adenocarcinomas necessitates further exploration of the unique clinical features and behavior of advanced-stage BRAF-mutant lung adenocarcinomas. METHODS: Data were reviewed for patients with advanced lung adenocarcinomas enrolled in the Lung Cancer Mutation Consortium whose tumors underwent testing for mutations in epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (KRAS), human epidermal growth factor receptor 2 (HER2), AKT1, BRAF, dual-specificity mitogen-activated protein kinase kinase 1 (MEK1), neuroblastoma RAS viral (v-ras) oncogene homolog (NRAS), and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α (PIK3CA); for anaplastic lymphoma kinase (ALK) translocations; and for MET amplification. RESULTS: Twenty-one BRAF mutations were identified in 951 patients with adenocarcinomas (2.2%; 95% confidence interval [CI], 1.4%-3.4%): 17 (81%; 95% CI, 60%-92%) were BRAF(V600E) mutations, and 4 were non-BRAF(V600E) mutations. Among the 733 cases tested for all 10 genes, BRAF mutations were more likely to occur than most other genotypic abnormalities in current or former smokers (BRAF vs sensitizing EGFR, 82% vs 36%, mid-P < .001; BRAF vs ALK, 39%, mid-P = .003; BRAF vs other mutations, 49%, mid-P = .02; BRAF vs patients with more than 1 oncogenic driver [doubleton], 46%, mid-P = .04.) The double-mutation rate was 16% among patients with BRAF mutations but 5% among patients with other genomic abnormalities (mid-P = .045). Differences were not found in survival between patients with BRAF mutations and those with other genomic abnormalities (P > .20). CONCLUSIONS: BRAF mutations occurred in 2.2% of advanced-stage lung adenocarcinomas, were most commonly V600E, and were associated with distinct clinicopathologic features in comparison with other genomic subtypes and with a high mutation rate in more than 1 gene. These findings underscore the importance of comprehensive genomic profiling in assessing patients with advanced lung adenocarcinomas.

Pulmonary symptoms measured by the national institutes of health lung score predict overall survival, nonrelapse mortality, and patient-reported outcomes in chronic graft-versus-host disease.

The 2005 National Institutes of Health (NIH) Consensus Conference recommended assessment of lung function in patients with chronic graft-versus-host disease (GVHD) by both pulmonary function tests (PFTs) and assessment of pulmonary symptoms. We tested whether pulmonary measures were associated with nonrelapse mortality (NRM), overall survival (OS), and patient-reported outcomes (PRO). Clinician and patient-reported data were collected serially in a prospective, multicenter, observational study. Available PFT data were abstracted. Cox regression models were fit for outcomes using a time-varying covariate model for lung function measures and adjusting for patient and transplantation characteristics and nonlung chronic GVHD severity. A total of 1591 visits (496 patients) were used in this analysis. The NIH symptom-based lung score was associated with NRM (P = .02), OS (P = .02), patient-reported symptoms (P < .001) and functional status (P < .001). Worsening of NIH symptom-based lung score over time was associated with higher NRM and lower survival. All other measures were not associated with OS or NRM; although, some were associated with patient-reported lung symptoms. In conclusion, the NIH symptom-based lung symptom score of 0 to 3 is associated with NRM, OS, and PRO measures in patients with chronic GVHD. Worsening of the NIH symptom-based lung score was associated with increased mortality.

Correlation between NIH composite skin score, patient-reported skin score, and outcome: results from the Chronic GVHD Consortium.

There are no validated criteria to measure skin response in chronic GVHD. In a prospectively assembled, multicenter cohort of patients with chronic GVHD (N = 458), we looked for correlation of change in several different scales recommended by the National Institutes of Health (NIH) Consensus with clinician and patient perception of change and overall survival. Of the clinician scales, the NIH composite 0-3 skin score was the only one that correlated with both clinician and patient perception of improvement or worsening. Of the patient-reported scales, the skin subscale of the Lee Symptom Scale was the only one that correlated with both clinician and patient perception of improvement or worsening. At study entry, NIH skin score 3 and Lee skin symptom score > 15 were both associated with worse overall survival. Worsening of NIH skin score at 6 months was associated with worse overall survival. Improvement in the Lee skin symptom score at 6 months was associated with improved overall survival. Our findings support the use of the NIH composite 0-3 skin score and the Lee skin symptom score as simple and sensitive measures to evaluate skin involvement in clinical trials as well as in the clinical monitoring of patients with cutaneous chronic GVHD.

Use of Response Permutation to Measure an Imaging Dataset's Susceptibility to Overfitting by Selected Standard Analysis Pipelines.

RATIONALE AND OBJECTIVES: This study demonstrates a method for quantifying the impact of overfitting on the receiving operator characteristic curve (AUC) when using standard analysis pipelines to develop imaging biomarkers. We illustrate the approach using two publicly available repositories of radiology and pathology images for breast cancer diagnosis. MATERIALS AND METHODS: For each dataset, we permuted the outcome (cancer diagnosis) values to eliminate any true association between imaging features and outcome. Seven types of classification models (logistic regression, linear discriminant analysis, Naïve Bayes, linear support vector machines, nonlinear support vector machine, random forest, and multi-layer perceptron) were fitted to each scrambled dataset and evaluated by each of four techniques (all data, hold-out, 10-fold cross-validation, and bootstrapping). After repeating this process for a total of 50 outcome permutations, we averaged the resulting AUCs. Any increase over a null AUC of 0.5 can be attributed to overfitting. RESULTS: Applying this approach and varying sample size and the number of imaging features, we found that failing to control for overfitting could result in near-perfect prediction (AUC near 1.0). Cross-validation offered greater protection against overfitting than the other evaluation techniques, and for most classification algorithms a sample size of at least 200 was required to assess as few as 10 features with less than 0.05 AUC inflation attributable to overfitting. CONCLUSION: This approach could be applied to any curated dataset to suggest the number of features and analysis approaches to limit overfitting.

Repeatability of 18F-FDG uptake in metastatic bone lesions of breast cancer patients and implications for accrual to clinical trials.

BACKGROUND: Standard measures of response such as Response Evaluation Criteria in Solid Tumors are ineffective for bone lesions, often making breast cancer patients that have bone-dominant metastases ineligible for clinical trials with potentially helpful therapies. In this study we prospectively evaluated the test-retest uptake variability of 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) in a cohort of breast cancer patients with bone-dominant metastases to determine response criteria. The thresholds for 95% specificity of change versus no-change were then applied to a second cohort of breast cancer patients with bone-dominant metastases. METHODS: For this study, nine patients with 38 bone lesions were imaged with 18F-FDG in the same calibrated scanner twice within 14 days. Tumor uptake was quantified by the most commonly used PET parameter, the maximum tumor voxel normalized by dose and body weight (SUVmax) and also by the mean of a 1-cc maximal uptake volume normalized by dose and lean-body-mass (SULpeak). The asymmetric repeatability coefficients with confidence intervals for SUVmax and SULpeak were used to determine the limits of 18F-FDG uptake variability. A second cohort of 28 breast cancer patients with bone-dominant metastases that had 146 metastatic bone lesions was imaged with 18F-FDG before and after standard-of-care therapy for response assessment. RESULTS: The mean relative difference of SUVmax and SULpeak in 38 bone tumors of the first cohort were 4.3% and 6.7%. The upper and lower asymmetric limits of the repeatability coefficient were 19.4% and - 16.3% for SUVmax, and 21.2% and - 17.5% for SULpeak. 18F-FDG repeatability coefficient confidence intervals resulted in the following patient stratification using SULpeak for the second patient cohort: 11-progressive disease, 5-stable disease, 7-partial response, and 1-complete response with three inevaluable patients. The asymmetric repeatability coefficients response criteria for SULpeak changed the status of 3 patients compared to the standard Positron Emission Tomography Response Criteria in Solid Tumors of ± 30% SULpeak. CONCLUSION: In evaluating bone tumor response for breast cancer patients with bone-dominant metastases using 18F-FDG SUVmax, the repeatability coefficients from test-retest studies show that reductions of more than 17% and increases of more than 20% are unlikely to be due to measurement variability. Serial 18F-FDG imaging in clinical trials investigating bone lesions in these patients, such as the ECOG-ACRIN EA1183 trial, benefit from confidence limits that allow interpretation of response.

Repeatability of 18F-FDG uptake in metastatic bone lesions of breast cancer patients and implications for accrual to clinical trials.

BACKGROUND: Standard measures of response such as Response Evaluation Criteria in Solid Tumors are ineffective for bone lesions, often making breast cancer patients with bone-dominant metastases ineligible for clinical trials with potentially helpful therapies. In this study we prospectively evaluated the test-retest uptake variability of 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) in a cohort of breast cancer patients with bone-dominant metastases to determine response criteria. The thresholds for 95% specificity of change versus no-change were then applied to a second cohort of breast cancer patients with bone-dominant metastases.In this study, nine patients with 38 bone lesions were imaged with 18F-FDG in the same calibrated scanner twice within 14 days. Tumor uptake was quantified as the maximum tumor voxel normalized by dose and body weight (SUVmax) and the mean of a 1-cc maximal uptake volume normalized by dose and lean-body-mass (SULpeak). The asymmetric repeatability coefficients with confidence intervals of SUVmax and SULpeak were used to determine limits of 18F-FDG uptake variability. A second cohort of 28 breast cancer patients with bone-dominant metastases that had 146 metastatic bone lesions was imaged with 18F-FDG before and after standard-of-care therapy for response assessment. RESULTS: The mean relative difference of SUVmax in 38 bone tumors of the first cohort was 4.3%. The upper and lower asymmetric limits of the repeatability coefficient were 19.4% and -16.3%, respectively. The 18F-FDG repeatability coefficient confidence intervals resulted in the following patient stratification for the second patient cohort: 11-progressive disease, 5-stable disease, 7-partial response, and 1-complete response with three inevaluable patients. The asymmetric repeatability coefficients response criteria changed the status of 3 patients compared to standard the standard Positron Emission Tomography Response Criteria in Solid Tumors of ±30% SULpeak. CONCLUSIONS: In evaluating bone tumor response for breast cancer patients with bone-dominant metastases using 18F-FDG uptake, the repeatability coefficients from test-retest studies show that reductions of more than 17% and increases of more than 20% are unlikely to be due to measurement variability. Serial 18F-FDG imaging in clinical trials investigating bone lesions from these patients, such as the ECOG-ACRIN EA1183 trial, benefit from confidence limits that allow interpretation of response.

Analysis of gastrointestinal and hepatic chronic graft-versus-host [corrected] disease manifestations on major outcomes: a chronic graft-versus-host [corrected] disease consortium study.

Although data support adverse prognosis of overlap subtype of chronic grant-versus-host disease (GVHD), the importance of site of gastrointestinal (GI) and type of hepatic involvement is not known. Using data from the Chronic GVHD Consortium observational cohort study (N = 567, total of 2115 visits), we examined whether the site of GI (esophageal, upper GI, or lower GI) and type of hepatic (bilirubin, alkaline phosphatase, alanine aminotransferase) involvement are associated with overall survival (OS) and nonrelapse mortality (NRM), symptoms, quality of life (QOL) and functional status measures. In multivariate analysis utilizing data from enrollment visits only, lower GI involvement (HR, 1.67; P = .05) and elevated bilirubin (HR, 2.46; P = .001) were associated with OS; both were also associated with NRM. In multivariable analysis using all visits (time-dependent covariates), GI score greater than zero (HR, 1.69; P = .02) and elevated bilirubin (HR, 3.73; P < .001) were associated with OS; results were similar for NRM. Any esophageal involvement and GI score greater than zero were associated with both symptoms and QOL, whereas elevated bilirubin was associated with QOL. We found no consistent evidence that upper GI involvement, alkaline phosphatase, alanine aminotransferase, or NIH liver score add prognostic value for survival, overall symptom burden, or QOL. These data support important differences in patient-reported outcomes according to GI and hepatic involvement among chronic GVHD-affected patients and identify those with elevated bilirubin or higher GI score at any time, or lower GI involvement at cohort enrollment, as patients at greater risk for mortality under current treatment approaches.

Multicenter trials using ¹8F-fluorodeoxyglucose (FDG) PET to predict chemotherapy response: effects of differential measurement error and bias on power calculations for unselected and enrichment designs.

BACKGROUND: Clinical validation of a predictive biomarker is especially difficult when the biomarker cannot be assessed retrospectively. A cost-effective, prospective multicenter replication study with rapid accrual is warranted prior to further validation studies such as a marker-based strategy for treatment selection. However, it is often unknown how measurement error and bias in a multicenter trial will differ from that in single-institution studies. PURPOSE: Power calculations using simulated data may inform the efficient design of a multicenter study to replicate single-institution findings. This case study used serial standardized uptake value (SUV) measures from (18)F-fluorodeoxyglucose (FDG) positron emission tomography (PET) to predict early response to breast cancer neoadjuvant chemotherapy. We examined the impact of accelerating accrual through increased inclusion of secondary sites with greater levels of measurement error and bias. We also examined whether enrichment designs based on breast cancer initial uptake could increase the study power for a fixed budget (200 total scans). METHODS: Reference FDG PET SUV data were selected with replacement from a single-institution trial; pathologic complete response (pCR) data were simulated using a logistic regression model predicting response by mid-therapy percent change in SUV. The impact of increased error for SUV measurements in multicenter trials was simulated by sampling from error and bias distributions: 20%-40% measurement error, 0%-40% bias, and fixed error/bias values. The proportion of patients recruited from secondary sites (with higher additional error/bias compared to primary sites) varied from 25% to 75%. RESULTS: Reference power (from source data with no added error) was 0.92 for N = 100 to detect an association between percentage change in SUV and response. With moderate (20%) simulated measurement error for 3/4, 1/2, and 1/4 of measurements and 40% for the remainder, power was 0.70, 0.61, and 0.53, respectively. Reduction of study power was similar for other manifestations of measurement error (bias as a percentage of true value, absolute error, and absolute bias). Enrichment designs, which recruit additional patients by not conducting a second scan in patients with unsuitable pre-therapy uptake (low baseline SUV), did not lead to greater power for studies constrained to the same total cost. LIMITATIONS: Simulation parameters could be incorrect, or not generalizable. Under a different logistic regression model relating mid-therapy percent change in SUV to pCR (with no relationship for patients with low baseline SUV, rather than the modest point estimate from reference data), the enrichment design did have somewhat greater power than the unselected design. CONCLUSION: Even moderate additional measurement error substantially reduced study power under both unselected and enrichment designs.

Summary: Appropriate Use Criteria for Estrogen Receptor-Targeted PET Imaging with 16α-18F-Fluoro-17ß-Fluoroestradiol.

PET imaging with 16α-18F-fluoro-17ß-fluoroestradiol (18F-FES), a radiolabeled form of estradiol, allows whole-body, noninvasive evaluation of estrogen receptor (ER). 18F-FES is approved by the U.S. Food and Drug Administration as a diagnostic agent "for the detection of ER-positive lesions as an adjunct to biopsy in patients with recurrent or metastatic breast cancer." The Society of Nuclear Medicine and Molecular Imaging (SNMMI) convened an expert work group to comprehensively review the published literature for 18F-FES PET in patients with ER-positive breast cancer and to establish appropriate use criteria (AUC). The findings and discussions of the SNMMI 18F-FES work group, including example clinical scenarios, were published in full in 2022 and are available at https://www.snmmi.org/auc Of the clinical scenarios evaluated, the work group concluded that the most appropriate uses of 18F-FES PET are to assess ER functionality when endocrine therapy is considered either at initial diagnosis of metastatic breast cancer or after progression of disease on endocrine therapy, the ER status of lesions that are difficult or dangerous to biopsy, and the ER status of lesions when other tests are inconclusive. These AUC are intended to enable appropriate clinical use of 18F-FES PET, more efficient approval of FES use by payers, and promotion of investigation into areas requiring further research. This summary includes the rationale, methodology, and main findings of the work group and refers the reader to the complete AUC document.

Validation of measurement scales in ocular graft-versus-host disease.

PURPOSE: To validate measurement scales for rating ocular chronic graft-versus-host disease (GVHD) after allogeneic hematopoietic cell transplantation. Candidate scales were recommended for use in clinical trials by the National Institutes of Health (NIH) Chronic GVHD Consensus Conference or have been previously validated in dry eye syndromes. DESIGN: Prospective follow-up study. PARTICIPANTS: Between August 2007 and June 2010, the study enrolled 387 patients with chronic GVHD in a multicenter, prospective, observational cohort. METHODS: Using anchor-based methods, we compared clinician or patient-reported changes in eye symptoms (8-point scale) with calculated changes in 5 candidate scales: The NIH eye score, patient-reported global rating of eye symptoms, Lee eye subscale, Ocular Surface Disease Index, and Schirmer test. Change was examined for 333 follow-up visits where both clinician and patient reported eye involvement at the previous visit. Linear mixed models were used to account for within-patient correlation. MAIN OUTCOME MEASURES: An 8-point scale of clinician or patient-reported symptom change was used as an anchor to measure symptom changes at the follow-up visits. RESULTS: In serial evaluations, agreement regarding improvement, stability, or worsening between the clinician and patient was fair (weighted kappa = 0.34). Despite only fair agreement between evaluators, all scales except the Schirmer test correlated with both clinician-reported and patient-reported changes in ocular GVHD activity. Among all scales, changes in the NIH eye scores showed the greatest sensitivity to symptom change reported by clinicians or patients. CONCLUSIONS: Our results support the use of the NIH eye score as a sensitive measure of eye symptom changes in clinical trials assessing treatment of chronic GVHD.

Sensitivity of changes in chronic graft-versus-host disease activity to changes in patient-reported quality of life: results from the Chronic Graft-versus-Host Disease Consortium.

BACKGROUND: The 2005 National Institute of Health Chronic Graft-versus-Host Disease Consensus Conference recommended collection of patient-reported outcomes in clinical trials on chronic graft-versus-host disease. We assessed whether changes in chronic graft-versus-host disease severity, determined using National Institute of Health criteria, clinicians' assessment or patients' self-evaluation, correlated with patient-reported quality of life as measured by the Short Form-36 and Functional Assessment of Cancer Therapy-Bone Marrow Transplant (FACT-BMT) instruments. DESIGN AND METHODS: Three-hundred and thirty-six adult patients (median age 52 years; range, 19-79) with chronic graft-versus-host disease from six transplant centers contributed baseline and follow-up data (from 936 visits overall). RESULTS: While the majority of the patients had stable chronic graft-versus-host disease, improvement or worsening was noted in approximately 40% of follow-up visits. Multivariable analysis demonstrated no association between change in chronic graft-versus-host disease severity evaluated by National Institute of Health criteria and change in quality of life, while clinician-reported changes in severity were associated with changes in some quality of life measures. Patient-reported changes in the severity of chronic graft-versus-host disease were associated with changes in all quality of life measures. Comparison of the Short Form-36 and the FACT-BMT suggested that the data collected in the Functional Assessment of Cancer Therapy-General (FACT-G) core survey are sufficient without the need for the Short Form-36 or the FACT-BMT subscale. CONCLUSIONS: We conclude that serial National Institute of Health and clinician-reported chronic graft-versus-host disease severity assessments cannot substitute for patient-reported outcomes in clinical trials. Collection of just the FACT-G instead of the Short Form-36 and the full FACT-BMT will decrease respondent burden without compromising quality of life assessment.

Probability of malignancy for lesions detected on breast MRI: a predictive model incorporating BI-RADS imaging features and patient characteristics.

OBJECTIVES: To predict the probability of malignancy for MRI-detected breast lesions with a multivariate model incorporating patient and lesion characteristics. METHODS: Retrospective review of 2565 breast MR examinations from 1/03-11/06. BI-RADS 3, 4 and 5 lesions initially detected on MRI for new cancer or high-risk screening were included and outcomes determined by imaging, biopsy or tumor registry linkage. Variables were indication for MRI, age, lesion size, BI-RADS lesion type and kinetics. Associations with malignancy were assessed using generalized estimating equations and lesion probabilities of malignancy were calculated. RESULTS: 855 lesions (155 malignant, 700 benign) were included. Strongest associations with malignancy were for kinetics (washout versus persistent; OR 4.2, 95% CI 2.5-7.1) and clinical indication (new cancer versus high-risk screening; OR 3.0, 95% CI 1.7-5.1). Also significant were age > = 50 years, size > = 10 mm and lesion-type mass. The most predictive model (AUC 0.70) incorporated indication, size and kinetics. The highest probability of malignancy (41.1%) was for lesions on MRI for new cancer, > = 10 mm with washout. The lowest (1.2%) was for lesions on high-risk screening, <10 mm with persistent kinetics. CONCLUSIONS: A multivariate model shows promise as a decision support tool in predicting malignancy for MRI-detected breast lesions.