Empirical Relative Biological Effectiveness (RBE) for Mandible Osteoradionecrosis (ORN) in Head and Neck Cancer Patients Treated With Pencil-Beam-Scanning Proton Therapy (PBSPT): A Retrospective, Case-Matched Cohort Study

Abstract

Purpose: To retrospectively investigate empirical relative biological effectiveness (RBE) for mandible osteoradionecrosis (ORN) in head and neck (H&N) cancer patients treated with pencil-beam-scanning proton therapy (PBSPT).

Methods: We included 1,266 H&N cancer patients, of which, 931 patients were treated with volumetric-modulated arc therapy (VMAT) and 335 were treated with PBSPT. Among them, 26 VMAT and 9 PBSPT patients experienced mandible ORN (ORN group), while all others were included in the control group. To minimize the impact of the possible imbalance in clinical factors between VMAT and PBSPT patients in the dosimetric comparison between these two modalities and the resulting RBE quantification, we formed a 1:1 case-matched patient cohort (335 VMAT patients and 335 PBSPT patients including both the ORN and control groups) using the greedy nearest neighbor matching of propensity scores. Mandible dosimetric metrics were extracted from the case-matched patient cohort and statistically tested to evaluate the association with mandibular ORN to derive dose volume constraints (DVCs) for VMAT and PBSPT, respectively. We sought the equivalent constraint doses for VMAT so that the critical volumes of VMAT were equal to those of PBSPT at different physical doses. Empirical RBEs of PBSPT for ORN were obtained by calculating the ratio between the derived equivalent constraint doses and physical doses of PBSPT. Bootstrapping was further used to get the confidence intervals.

Results: Clinical variables of age, gender, tumor stage, prescription dose, chemotherapy, hypertension or diabetes, dental extraction, smoking history, or current smoker were not statistically related to the incidence of ORN in the overall patient cohort. Smoking history was found to be significantly associated with the ORN incidence in PBSPT patients only. V40Gy[RBE], V50Gy[RBE], and V60Gy[RBE] were statistically different (p<0.05) between the ORN and control group for VMAT and PBSPT. Empirical RBEs of 1.58(95%CI: 1.34-1.64), 1.34(95%CI: 1.23-1.40), and 1.24(95%: 1.15-1.26) were obtained for proton dose at 40 Gy[RBE=1.1], 50 Gy[RBE=1.1] and 60 Gy[RBE=1.1], respectively.

Conclusions: Our study suggested that RBEs were larger than 1.1 at moderate doses (between 40 and 60 Gy[RBE=1.1]) with high LET for mandible ORN. RBEs are underestimated in current clinical practice in PBSPT. The derived DVCs can be used for PBSPT plan evaluation and optimization to minimize the incidence rate of mandible ORN.

Keywords: dose LET volume histogram (DLVH); head and neck (H&N) cancer; linear energy transfer (LET); mandibular osteoradionecrosis; pencil beam scanning proton therapy (PBSPT); relative biological effectiveness; volume modulated arc-therapy (VMAT).

Figure 1

Dose-LET volume histogram (DLVH) of mandible. (A) DLVHs of mandible for a typical patient with ORN and (B) without ORN. DL5%, DL20%, and DL50% lines were plotted with different colors. Red dashed ovals indicate an area containing voxels with moderate doses but high LETs for this ORN patient.

Figure 2

Dose and LET distributions of representative patients with and without osteoradionecrosis (ORN) treated with volumetric-modulated arc therapy (VMAT) and pencil-beam-scanning proton therapy (PBSPT) (A) Axial (top row) and sagittal (bottom row) view of dose distributions of typical patients treated with VMAT (left two columns) and PBSPT (right two columns) with (1^st and 3^rd column) and without (2^nd and 4^th column) ORN. (B) Axial (top row) and sagittal (bottom row) view of LET distributions of typical patients treated with PBSPT with (left column) and without (right column) ORN. Contour colors: pink: ORN injury regions; cyan: CTV_high; blue: CTV_low; green: mandible.

Figure 3

Box plots of volume value distributions of DVH indices. (A) Box plots of volume value distributions of V40Gy[RBE], V50Gy[RBE], V60Gy[RBE], V70Gy[RBE], and V75Gy[RBE] for all photon (blue) and proton (red) patients. (B) Box plots of volume value distributions of V40Gy[RBE], V50Gy[RBE], V60Gy[RBE], V70Gy[RBE], and V75Gy[RBE] for patients with (dark color) and without (light color) ORN treated with VMAT (blue) and PBSPT (red). Blue and red horizontal lines in (A) and (B) indicated the derived DVCs for VMAT and PBSPT, respectively. (C) Volume tolerance curves for VMAT (blue) and PBSPT (red) based on the derived DVCs. Gray circle indicates the position at the intersection between the VMAT volume tolerance curve and a horizontal line (indicated by the gray arrow) with the same critical volume value as the corresponding DVC of the PBSPT volume tolerance curve. The corresponding dose at the gray circle was the equivalent constraint dose. Error bars indicate the 95% confidence intervals obtained by a bootstrapping of 1,000 times.