KEY POINTS
- This single-center retrospective study included 200 patients with nasopharyngeal carcinoma in two sequential cohorts of 100. The first cohort established statistical process-control benchmarks, while the second used those centerlines and control limits to guide plan optimization.
- All patients received nine-field coplanar IMRT planned in Pinnacle by the same senior physicist. Prescription doses were 70 Gy in 33 fractions to PGTVnx and PGTVnd, 62 Gy in 33 fractions to PTV1, and 54 Gy in 33 fractions to PTV2, with a 3-mm planning margin.
- After implementation of statistical process control, mean laryngeal dose fell from 33.54 to 23.51 Gy, an absolute reduction of 10.03 Gy. Mean parotid dose decreased from 35.68 to 33.22 Gy, a 2.46-Gy reduction, and parotid D50 decreased from 31.63 to 28.58 Gy, a 3.05-Gy reduction. All three differences were p < 0.001.
- Target coverage was preserved. D95 for PGTVnx, PGTVnd, PTV1, and PTV2 showed no significant difference between the two cohorts, all p > 0.05.
- Process capability improved modestly after implementation. Cpk increased from 1.43 to 1.45 for laryngeal mean dose, 1.04 to 1.11 for parotid mean dose, and 1.07 to 1.13 for parotid D50.
- All second-phase OAR measurements remained within the derived statistical control limits, supporting improved process stability. However, both parotid Cpk values remained within the authors' 1.00-1.33 "normal but requiring attention" range, rather than reaching the higher capability categories.
- The study was sequential rather than randomized and used a single center, one senior planner, and only larynx and parotids as monitored OARs. Planner learning effects cannot be separated from the effect of statistical process control, and no xerostomia, dysphagia, or other clinical toxicity outcomes were measured.
CLINICAL TAKEAWAY
Statistical process control may provide a simple, TPS-independent way to turn institutional planning performance into continuously updated optimization benchmarks. The dosimetric gains are substantial, particularly the 10-Gy reduction in mean laryngeal dose, but this remains a single-center planning study and cannot establish improved patient outcomes.