Radiotherapy, including brachytherapy, is a major therapeutic regimen for cervical cancer. Radioresistance is a decisive factor in radiation treatment failure. Tumor-associated macrophages (TAMs) and cancer-associated fibroblasts (CAFs) in the tumor microenvironment are critical factors in the curative effects of cancer therapies. However, the interactions between TAMs and CAFs in the context of ionizing radiation are not fully understood. This study was undertaken to investigate whether M2 macrophages induce radioresistance in cervical cancer and to explore the TAMs’ phenotypic transformation after IR and its underlying mechanisms. The radioresistance of cervical cancer cells was enhanced after being co-cultured with M2 macrophages. TAMs tended to undergo M2 polarization after high-dose irradiation, which was strongly associated with CAFs in both mouse models and patients with cervical cancer. Additionally, cytokine and chemokine analysis was performed to find that high-dose irradiated CAFs promoted macrophage polarization towards the M2 phenotype through chemokine (C-C motif) ligand 2. Collectively, our results highlight the crucial role that high-dose irradiated CAFs play in the regulation of M2 phenotype polarization, which ultimately induces radioresistance in cervical cancer.

Objective. To evaluate the feasibility of using a deep learning dose prediction approach to identify patients who could benefit most from proton therapy based on the normal tissue complication probability (NTCP) model. Approach. Two 3D UNets were established to predict photon and proton doses. A dataset of 95 patients with localized prostate cancer was randomly partitioned into 55, 10, and 30 for training, validation, and testing, respectively. We selected NTCP models for late rectum bleeding and acute urinary urgency of grade 2 or higher to quantify the benefit of proton therapy. Propagated uncertainties of predicted ΔNTCPs resulting from the dose prediction errors were calculated. Patient selection accuracies for a single endpoint and a composite evaluation were assessed under different ΔNTCP thresholds. Main results. Our deep learning-based dose prediction technique can reduce the time spent on plan comparison from approximately 2 days to as little as 5 seconds. The expanded uncertainty of predicted ΔNTCPs for rectum and bladder endpoints propagated from the dose prediction error were 0.0042 and 0.0016, respectively, which is less than one-third of the acceptable tolerance. The averaged selection accuracies for rectum bleeding, urinary urgency, and composite evaluation were 90%, 93.5%, and 93.5%, respectively. Significance. Our study demonstrates that deep learning dose prediction and NTCP evaluation scheme could distinguish the NTCP differences between photon and proton treatment modalities. In addition, the dose prediction uncertainty does not significantly influence the decision accuracy of NTCP-based patient selection for proton therapy. Therefore, automated deep learning dose prediction and NTCP evaluation schemes can potentially be used to screen large patient populations and to avoid unnecessary delays in the start of prostate cancer radiotherapy in the future.

Objective. Monoenergetic high-energy Bragg peak (monoBP) proton therapy has emerged as promising candidates for conformal FLASH radiotherapy (FLASH-RT). However, the beam-specific proton modulation devices are needed for dose conformality. Meanwhile, the beam switching time could disrupt ultra-high dose rate (UHDR), thereby compromising the FLASH effect. This study aims to propose a novel monoBP conformal FLASH-RT methodology with the only utilization of universal range shifter (URS). Approach. An optimization algorithm, which optimized both URS thickness and dose distribution, was implemented based on multi-field optimization. Two proton FLASH techniques with 218 MeV proton beams were investigated for 10 brain and 10 lung cancer cases: (1) URS-modulated monoBP FLASH plan, and (2) transmission beam (TB) FLASH plan. All plans were optimized under the same optimization constraints with robust scenarios. Delivery parameters, dose, and dose rate metrics of the two plans were analyzed and compared. Main Results. Both modalities achieved similar dose coverage, with D _98% of target meeting the clinical requirement for brain (BP: 98.2 ± 0.7%, TB: 98.6 ± 0.5%) and lung (BP: 97.8 ± 0.3%, TB: 98.7 ± 0.5%) cancer cases. However, compared to TB strategy, the BP FLASH plans reduced the D _mean of normal tissue for brain (BP: 16.6 ± 3.1 Gy, TB: 19.7 ± 5.4 Gy, p < 0.05) and lung (BP: 3.3 ± 1.7 Gy, TB: 4.2 ± 1.8 Gy, p < 0.05) cases. For dose-averaged dose rate, the V _{40Gy/s_DADR} of both modalities reached 100% in brain and lung cancer cases; while for averaged dose rate, the V _{40Gy/s_ADR} was 37.4 % and 26.6 % for BP plans in brain and lung cancer cases, respectively, and was 61.5% and 61.3% for TB plans in brain and lung cancer cases, respectively. Significance. A monoBP conformal FLASH-RT methodology was proposed, utilizing only a fixed URS. The proposed monoBP FLASH-RT reached the requirement of FLASH effect, while demonstrating better dose protection, compared to the TB FLASH-RT.