Concordance Between Institutional and Central Pathology, Surgery, and Radiology Reviews on the Children's Oncology...

The Children's Oncology Group (COG) AREN03B2 Renal Tumor Biology/Classification Study undertook real-time central reviews to facilitate accurate diagnosis, staging, and risk stratification for enrollment on therapeutic trials. This study assessed concordance between central and institutional reviews to inform future clinical trials and real-world practice.

Eligible patients who enrolled on AREN03B2 from 2006 to 2019 were included. COG experts performed real-time central radiology review to assess lung metastasis and bilateral renal lesions, pathology review for histology and pathological stage, and surgery review of local staging. An initial risk assignment (IRA) was made based on these reviews. Concordance between institutional and central reviews was measured using raw rates of agreement (%) and Cohen's kappa coefficient for inter-rater reliability.

6311 patients were included. Raw rate of agreement between institutional and central pathological diagnosis was 91.1% (Cohen's kappa: 0.79 [95% CI: 0.78-0.81]). Among patients with unilateral tumors, only 181/269 (67%) with diffuse anaplastic Wilms tumor were correctly identified by local institutions. Raw rate of agreement between assessments of computed tomography (CT) scans for lung nodules was 96.1% (Cohen's kappa: 0.88 [95% CI: 0.86-0.89]). Raw rate of agreement between surgical reviews for local stage was 90% (Cohen's kappa: 0.84 [95% CI: 0.83-0.85]). Raw rate of agreement between calculated institutional and central IRA was 87% (Cohen's kappa: 0.84 [95% CI: 0.83-0.85]).

Central reviews were highly concordant with institutional assessments, though lack of recognition of anaplastic histology by institutional pathologists was notable. Institutions should consider expert pathology consultation in the absence of a clinical trial with integrated central review.