The value of standards for health datasets in artificial intelligence-based applications.

Artificial intelligence as a medical device is increasingly being applied to healthcare for diagnosis, risk stratification and resource allocation. However, a growing body of evidence has highlighted the risk of algorithmic bias, which may perpetuate existing health inequity. This problem arises in part because of systemic inequalities in dataset curation, unequal opportunity to participate in research and inequalities of access. This study aims to explore existing standards, frameworks and best practices for ensuring adequate data diversity in health datasets. Exploring the body of existing literature and expert views is an important step towards the development of consensus-based guidelines. The study comprises two parts: a systematic review of existing standards, frameworks and best practices for healthcare datasets; and a survey and thematic analysis of stakeholder views of bias, health equity and best practices for artificial intelligence as a medical device. We found that the need for dataset diversity was well described in literature, and experts generally favored the development of a robust set of guidelines, but there were mixed views about how these could be implemented practically. The outputs of this study will be used to inform the development of standards for transparency of data diversity in health datasets (the STANDING Together initiative).

Do Cell-Cycle Phase-Specific Markers Predict Disease Grade, Stage, and Outcome in Cervical Carcinoma?

AIMS: Multiparameter analysis of cell cycle markers has shown a strong relationship between cell cycle progression and tumor grade, stage, and clinical outcome in penile, renal, ovarian, and breast cancers. We sought to link expression of cell cycle phase-specific markers in cervical cancer to tumor grade, stage, and clinical outcome to investigate their potential use as prognostic and predictive markers. METHODS: Pretreatment biopsy material was obtained from 35 patients with cervical cancer (stage IB2-IVA) and 12 normal cervix control cases. Each patient was treated with neoadjuvant chemotherapy followed by chemoradiation. Immunohistochemical staining was performed using a panel of cell cycle phase markers: replication licensing factors: Mcm2 (minichromosome maintenance 2) and geminin, and the standard proliferation marker Ki67 (clone MIB-1). RESULTS: The expression levels of each cell cycle biomarker were very high in all cases of squamous cell carcinoma of the cervix regardless of grade or stage of disease. In our cohort, all cases displayed an aggressive, so-called actively cycling phenotype. Univariate analysis showed that none of the cell cycle biomarkers predicted grade, stage, or clinical outcome. CONCLUSIONS: Cell cycle phase-specific markers do not appear to predict disease grade, stage, or outcome in our sample of patients with cervical cancer. This is not surprising, given that the expression of each cell cycle biomarker was very high in all cases.Indeed, all the cases of squamous cell carcinoma of the cervix (n = 28) and all but 1 of the adenocarcinomas (n = 7) in this study displayed an aggressive "actively cycling" phenotype. This predominance of actively cycling tumors is unusual and may reflect the viral etiology underlying the disease. These preliminary findings raise many interesting questions including the prognostic value of disease grade and markers of proliferation in cervical tumors as reliable prognostic indicators. Further work on a larger cohort of patients is warranted.

Adjuvant therapy in stage III endometrial cancer: treatment outcomes and survival. a single-institution retrospective study.

OBJECTIVE: As adjuvant treatment of advanced-stage endometrial cancer remains undefined, we sought to review and describe the outcomes of patients with International Federation of Obstetrics and Gynecology stage III endometrial cancer treated with chemotherapy and/or radiotherapy after primary surgery. METHODS: We conducted a retrospective cohort study of patients with stage III disease treated at University College London Hospitals from 2002 to 2009. Patients were eligible if they received adjuvant treatment at our center. We excluded those with any synchronous gynecologic tumor and patients who underwent surgery but not adjuvant treatment at the center. RESULTS: Stages IIIA, IIIB, and IIIC tumors accounted for 60%, 10%, and 30%, respectively. The median age was 67 years (range, 37-94 years). Sixty-five percent were pure endometrioid tumors, and 65% were high-grade (grade 3) tumors. Eighty-one patients received adjuvant treatment, 9% received chemotherapy alone, 28% received radiotherapy alone, and 63% received sequential combined chemotherapy followed by external beam radiotherapy with vaginal vault brachytherapy. In multivariate analysis, there was a significant difference between the adjuvant treatment groups for disease-free survival (DFS) and overall survival (OS) with those who received chemotherapy (DFS: P = 0.0001; hazard ratio [HR], 6.2; 95% confidence interval [CI], 2.47-15.8; OS: P = 0.003; HR, 6.0; CI, 2.2-16.6) or radiotherapy alone (DFS: P = 0.06; HR, 1.88; CI, 0.97-3.7; OS: P = 0.025; HR, 2.1; CI, 1.1-4.1) having a poorer survival compared to combined treatment. Overall survival at 3 years and 5 years were 57% and 47%, respectively, for all 81 patients who received any adjuvant treatment. CONCLUSIONS: Sequential combined adjuvant chemotherapy and radiotherapy may be associated with a significant improvement in survival compared with chemotherapy or radiotherapy alone. Univariate and multivariate analysis showed that advanced age, high grade, and presence of lymphovascular space invasion were associated with poor DFS and OS. For patients with documented recurrence (n = 41), there was no clear relationship between site of recurrence and type of adjuvant treatment given.