Continuing Education Needs Identified During the Initial Experience of Delivering 3D Conformal Radiotherapy for Cervical Cancer Patients in Tanzania

Abstract

In resource-limited settings where experience with 3D conformal radiation therapy (3DCRT) is still growing, review of simulation CTs and radiation therapy plans may illuminate topics for continuing education. We conducted a review of an early cohort of cervical cancer patients treated with 3DCRT in Tanzania to identify education needs that may be generalizable to similar resource-limited settings.In Tanzania, 3DCRT capabilities began in August 2018. A retrospective review of 30 consecutive patients who received definitive whole pelvis 3DCRT for cervical cancer from 2019 to 2020 was conducted after IRB approval. Each patient's simulation CT was evaluated for evidence of advanced disease not detected on initial diagnostic work-up, including pelvic or para-aortic lymphadenopathy (LAD) defined as axial dimension > = 1 cm. An independent set of contours was generated according to United States national consensus guidelines. DICE similarity coefficient (DSC) agreement analysis was conducted between contours. As non-systematic variations in planning target volume (PTV) labelling, summations, and expansions precluded direct comparisons, only clinical target volume (CTV) contours were compared in this analysis. Clinically treated 3DCRT plans were evaluated using original and consensus-based contours, and dosimetric coverage was compared using Wilcoxon signed-rank testing.Simulation CT was the sole means of pre-treatment cross-sectional abdominopelvic imaging in 24/30 patients (80%). The pre-simulation clinical stages were FIGO IB (n = 6), IIA (n = 10), IIB (n = 12), IIIA (n = 1), and IVA (n = 1). Simulation CT upstaged 20 of 30 patients (66.7%) to IIIC1 or IIIC2 disease, identifying eight patients with para-aortic LAD and 13 patients with pelvic LAD. Agreement analysis between clinical and consensus-based contours revealed high agreement in the primary CTV (DSC median 0.83, IQR 0.68 - 0.87), bladder (0.93, IQR 0.88 - 0.95), and femur (0.86) contours, but lower agreement in lymph node CTV (CTVn) contours (DSC 0.59, IQR 0.46 - 0.67). The median volume of clinical-CTVn covered by 45 Gy was 99.7% (IQR 97.5 - 100) while the median volume of consensus-CTVn covered by 45 Gy was 83.3% (IQR 80.1 - 90.1) (P < 0.001). The most frequently undercovered regions were the superior and inferior pre-sacral nodal regions.In resource-limited settings, early case review after 3DCRT initiation may reveal key elements in radiation planning that can guide continuing education topics. In our setting, high yield interventions include systematic evaluation of pelvic and para-aortic LAD on simulation CTs, educational refreshers about nodal contouring, and development of resource-appropriate treatment paradigms when advanced nodal disease is detected.