J Appl Clin Med Phys. 2026 May;27(5):e70635. doi: 10.1002/acm2.70635.
ABSTRACT
BACKGROUND: Conformity evaluation is essential for assessing radiotherapy plan quality, yet widely used indices such as the ICRU and Paddick Conformity Index do not distinguish between target underdosage and peripheral overdose. As clinical protocols differ in their target coverage definitions ICRU 50/62, ICRU 83, and RTOG/NRG direct comparisons of plan conformity across protocols remain challenging. Recently proposed metrics the Universal Conformity Index (CI) and Unconformity Indices (UCIu for underdose and UCIo for overdose) provide a decomposed, interpretable assessment of dose target relationships. However, protocol-specific optimal ranges for these indices and their clinical achievability have not been systematically established for pelvic radiotherapy.
PURPOSE: To determine protocol-specific optimal (p10-p90) ranges for CI, UCIu, and UCIo across pelvic sites under ICRU 50/62, ICRU 83, and RTOG/NRG frameworks, and to verify whether these optimal values are clinically achievable using modern IMRT plans.
METHODS: This study analyzed 40 anonymized pelvic cancer cases (20 prostate, 10 rectum, 10 endometrium). All cases were replanned under standardized dose prescriptions following ICRU 50/62, ICRU 83, and RTOG/NRG protocols. Two geometric configurations were evaluated: Target 1 (strict) and Target 2 (conservative). Stage 1: Using protocol-specific V95(%) definitions, theoretical optimal CI, UCIu, and UCIo bands (p10-p90) were calculated separately for each site and protocol under Target 1 and Target 2. Stage 2: All 40 cases were combined into a pooled pelvic dataset to derive region-wide optimal ranges and assess generalizability across sites. Stage 3: Clinical IMRT plans satisfying ICRU, RTOG, and QUANTEC constraints were generated for all patients, and clinical CI, UCIu, and UCIo values were compared against theoretical optimal bands. Statistical analysis included percentile bands, 95% Confidence Interval, p value, and effect size (η2).
RESULTS: Across all pelvic sites, CI values decreased progressively from ICRU 50/62 to ICRU 83 and RTOG/NRG (p < 0.001), while UCIu and UCIo increased systematically, reflecting protocol-driven differences in coverage tolerance. Margin expansion from Target 1 to Target 2 further reduced CI and increased UCIo. Pooled pelvic optimal ranges showed tight reproducibility with narrow 95% confidence intervals (± 0.006-0.016). Strong protocol effects were observed for CI (η2 = 0.259-0.519), UCIu (η2 = 0.997), and UCIo (η2 up to 0.182). Clinically generated IMRT plans demonstrated high-quality target coverage (V95% ≥ 98.9%) and acceptable OAR doses. Clinical CI, UCIu, and UCIo values closely matched the optimal p10-p90 bands. Prostate and rectum plans aligned fully with all protocol-specific optimal ranges, while endometrium plans matched RTOG/NRG and Target-2 thresholds, with minor deviations under strict Target-1 geometry.
CONCLUSIONS: This study provides the protocol-specific and clinically achievable optimal ranges for CI, UCIu, and UCIo for pelvic radiotherapy across ICRU 50/62, ICRU 83, and RTOG/NRG frameworks. The combined theoretical, pooled, and clinical validation approach demonstrates that these decomposed conformity indices are robust, reproducible, and directly applicable to routine IMRT plan evaluation. The resulting optimal bands offer standardized benchmarks for assessing dose conformity and dose spill, improving cross-protocol comparability and supporting future development of quantitative conformity-based guidelines in radiotherapy.
PMID:42178514 | DOI:10.1002/acm2.70635
