Dose–effect response in image-guided adaptive brachytherapy for cervical cancer: A systematic review and meta-regression analysis

Abstract

PurposeThis study aimed to integrate and update the dose–effect relationship between volumetric dose and local control for cervical cancer brachytherapy. Methods and MaterialsWe identified studies that reported high-risk clinical target volume (HR-CTV) D90 and local control probability by searching PubMed, Web of Science, and the Cochrane Library databases through Oct 27, 2019. The regression analyses were performed using a probit model between HR-CTV D90, D100, intermediate-risk clinical target volume (IR-CTV) D90, and dose to Point A vs. local control probability. Subgroup analyses were performed according to stratification: time of local control, income level of the country or region, stage of cancer, pathology, mean volume of HR-CTV, dose rate, image modality, concurrent chemoradiotherapy proportion, interstitial proportion, and mean overall treatment time. ResultsThirty-three studies encompassing 2893 patients were included. The probit model showed a significant relationship between the HR-CTV D90 value and the local control probability, p < 0.0001. The D90 corresponding to a probability of 90% local control was 83.7 GyEQD2,10 (80.6–87.8 GyEQD2,10). Of the 33 studies included in our analysis, eight studies, including 1172 patients, reported the IR-CTV D90 value, ranging from 59.1 GyEQD2,10 to 72.3 GyEQD2,10. The probit model also showed a significant relationship between the IR-CTV D90 value and the local control probability, p = 0.0464. The 60 GyEQD2,10 for IR-CTV D90 corresponded to an 86.1% local control probability (82.0%–89.8%). ConclusionsA significant dependence of local control on HR-CTV D90 and IR-CTV D90 was found. A tumor control probability of >90% can be expected at doses >84 GyEQD2,10 and 69 GyEQD2,10, respectively, based on an updated meta-regression analysis.