RESUMO
Purpose: Breast ultrasound suffers from low positive predictive value and specificity. Artificial intelligence (AI) proposes to improve accuracy, reduce false negatives, reduce inter- and intra-observer variability and decrease the rate of benign biopsies. Perpetuating racial/ethnic disparities in healthcare and patient outcome is a potential risk when incorporating AI-based models into clinical practice; therefore, it is necessary to validate its non-bias before clinical use. Approach: Our retrospective review assesses whether our AI decision support (DS) system demonstrates racial/ethnic bias by evaluating its performance on 1810 biopsy proven cases from nine breast imaging facilities within our health system from January 1, 2018 to October 28, 2021. Patient age, gender, race/ethnicity, AI DS output, and pathology results were obtained. Results: Significant differences in breast pathology incidence were seen across different racial and ethnic groups. Stratified analysis showed that the difference in output by our AI DS system was due to underlying differences in pathology incidence for our specific cohort and did not demonstrate statistically significant bias in output among race/ethnic groups, suggesting similar effectiveness of our AI DS system among different races (p>0.05 for all). Conclusions: Our study shows promise that an AI DS system may serve as a valuable second opinion in the detection of breast cancer on diagnostic ultrasound without significant racial or ethnic bias. AI tools are not meant to replace the radiologist, but rather to aid in screening and diagnosis without perpetuating racial/ethnic disparities.
RESUMO
PURPOSE: To assess response to programmed death-1 (PD-1) monotherapy (nivolumab) in hepatocellular carcinoma (HCC) patients using RECIST1.1, modified RECIST (mRECIST), and immune RECIST (iRECIST). A secondary objective was to identify clinicolaboratory and imaging variables predictive of progressive disease (PD) and overall survival (OS). METHODS: Patients with HCC treated with nivolumab at a single institution from 5/2016 to 12/2019 with MRI or CT performed ≥ 4 weeks post treatment were retrospectively assessed. Patients who received concurrent locoregional, radiation, or other systemic therapies were excluded. Response was assessed by 2 observers in consensus using RECIST1.1, mRECIST, and iRECIST at 3/6/9/12-month time points. Time to progression (TTP) and OS were recorded. Clinicolaboratory and imaging variables were evaluated as predictors of PD and OS using uni-/multivariable and Cox regression analyses. RESULTS: Fifty-eight patients (42M/16F) were included. 118 target lesions (TL) were identified before treatment. Baseline mean TL size was 49.1 ± 43.5 mm (range 10-189 mm) for RECIST1.1/iRECIST and 46.3 ± 42.3 mm (range 10-189 mm) for mRECIST. Objective response rate (ORR) was 21% for mRECIST/iRECIST/RECIST1.1, with no cases of pseudoprogression. Median OS and median TTP were 717 days and 127 days for RECIST1.1/mRECIST/iRECIST-iUPD (unconfirmed PD). Older age, MELD/Child-Pugh scores, AFP, prior transarterial radioembolization (TARE), and larger TL size were predictive of PD and/or poor OS using mRECIST/iRECIST. The strongest predictor of PD (HR = 2.49, 95% CI 1.29-4.81, p = 0.007) was TARE. The strongest predictor of poor OS was PD by mRECIST/iRECIST at 3 months (HR = 2.26, 95% CI 1.00-5.10, p = 0.05) with borderline significance. CONCLUSION: Our results show ORR of 21%, equivalent for mRECIST, iRECIST, and RECIST1.1 in patients with advanced HCC clinically treated with nivolumab.