Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics (AIFB) · 2026
Logs are all you need: Autonomous detection of LINAC monitor chamber drift.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Reynard Dimitri, Gastine Maël, Mazzara Christophe
Аннотация
BACKGROUND AND PURPOSE: To assess whether monitor-chamber sensitivity drift can be detected directly from Varian® Halcyon Trajectory Logs (T-logs), enabling autonomous, sensor-free output quality assurance (QA).
METHODS: Internal T-logs data sampled every 20 ms were analyzed to compare expected versus actual cumulative monitor units (MU). A pulsed-beam model explained sampling-induced quantization appearing as a high-frequency sawtooth in the error signal. A two-parameter analytic sawtooth template (frequency, amplitude) was phase-aligned by cross-correlation and subtracted to remove this artifact. The filtered residual was summarized per daily output test (median) to form drift indicators along the dose-odometer (DoD). Modeled versus experimental spectra were compared using spectral RMSE, NRMSE, Pearson correlation, median peak error, and band-power differences at 1, 16, and 17 Hz. Between recalibrations, piecewise linear regressions of log-derived and external Daily QA outputs were compared via segment-wise ANCOVA. Recalibrations were finally identified directly from T-logs using a Generalized Likelihood Ratio (GLR) change-point test.
RESULTS: Modeled and experimental error spectra showed strong agreement (RMSE ≈ 1-2 dB; NRMSE ≈ 0.03-0.04; correlation ≈ 0.97-0.98). ANCOVA confirmed similar drift slopes between log-derived and Daily QA outputs in most intervals; early discrepancies were mainly intercept shifts. GLR successfully detected recent recalibrations on both machines without false alarms.
CONCLUSIONS: T-logs embed a reproducible signature of chamber sensitivity drift. A lightweight two-parameter sawtooth filter enables immediate artifact suppression and autonomous calibration tracking, complementing existing QA and paving the way toward predictive, log-based maintenance.
METHODS: Internal T-logs data sampled every 20 ms were analyzed to compare expected versus actual cumulative monitor units (MU). A pulsed-beam model explained sampling-induced quantization appearing as a high-frequency sawtooth in the error signal. A two-parameter analytic sawtooth template (frequency, amplitude) was phase-aligned by cross-correlation and subtracted to remove this artifact. The filtered residual was summarized per daily output test (median) to form drift indicators along the dose-odometer (DoD). Modeled versus experimental spectra were compared using spectral RMSE, NRMSE, Pearson correlation, median peak error, and band-power differences at 1, 16, and 17 Hz. Between recalibrations, piecewise linear regressions of log-derived and external Daily QA outputs were compared via segment-wise ANCOVA. Recalibrations were finally identified directly from T-logs using a Generalized Likelihood Ratio (GLR) change-point test.
RESULTS: Modeled and experimental error spectra showed strong agreement (RMSE ≈ 1-2 dB; NRMSE ≈ 0.03-0.04; correlation ≈ 0.97-0.98). ANCOVA confirmed similar drift slopes between log-derived and Daily QA outputs in most intervals; early discrepancies were mainly intercept shifts. GLR successfully detected recent recalibrations on both machines without false alarms.
CONCLUSIONS: T-logs embed a reproducible signature of chamber sensitivity drift. A lightweight two-parameter sawtooth filter enables immediate artifact suppression and autonomous calibration tracking, complementing existing QA and paving the way toward predictive, log-based maintenance.