Multimodal radiotherapy dose prediction using a multi-task deep learning model.

BACKGROUND: In radiation therapy (RT), accelerated partial breast irradiation (APBI) has emerged as an increasingly preferred treatment modality over conventional whole breast irradiation due to its targeted dose delivery and shorter course of treatment. APBI can be delivered through various modalities including Cobalt-60-based systems and linear accelerators with C-arm, O-ring, or robotic arm design. Each modality possesses distinct features, such as beam energy or the degrees of freedom in treatment planning, which influence their respective dose distributions. These modality-specific considerations emphasize the need for a quantitative approach in determining the optimal dose delivery modality on a patient-specific basis. However, manually generating treatment plans for each modality across every patient is time-consuming and clinically impractical. PURPOSE: We aim to develop an efficient and personalized approach for determining the optimal RT modality for APBI by training predictive models using two different deep learning-based convolutional neural networks. The baseline network performs a single-task (ST), predicting dose for a single modality. Our proposed multi-task (MT) network, which is capable of leveraging shared information among different tasks, can concurrently predict dose distributions for various RT modalities. Utilizing patient-specific input data, such as a patient's computed tomography (CT) scan and treatment protocol dosimetric goals, the MT model predicts patient-specific dose distributions across all trained modalities. These dose distributions provide patients and clinicians quantitative insights, facilitating informed and personalized modality comparison prior to treatment planning. METHODS: The dataset, comprising 28 APBI patients and their 92 treatment plans, was partitioned into training, validation, and test subsets. Eight patients were dedicated to the test subset, leaving 68 treatment plans across 20 patients to divide between the training and validation subsets. ST models were trained for each modality, and one MT model was trained to predict doses for all modalities simultaneously. Model performance was evaluated across the test dataset in terms of Mean Absolute Percent Error (MAPE). We conducted statistical analysis of model performance using the two-tailed Wilcoxon signed-rank test. RESULTS: Training times for five ST models ranged from 255 to 430 min per modality, totaling 1925 min, while the MT model required 2384 min. MT model prediction required an average of 1.82 s per patient, compared to ST model predictions at 0.93 s per modality. The MT model yielded MAPE of 1.1033 ± 0.3627% as opposed to the collective MAPE of 1.2386 ± 0.3872% from ST models, and the differences were statistically significant (p = 0.0003, 95% confidence interval = [-0.0865, -0.0712]). CONCLUSION: Our study highlights the potential benefits of a MT learning framework in predicting RT dose distributions across various modalities without notable compromises. This MT architecture approach offers several advantages, such as flexibility, scalability, and streamlined model management, making it an appealing solution for clinical deployment. With such a MT model, patients can make more informed treatment decisions, physicians gain more quantitative insight for pre-treatment decision-making, and clinics can better optimize resource allocation. With our proposed goal array and MT framework, we aim to expand this work to a site-agnostic dose prediction model, enhancing its generalizability and applicability.

Comparison of first-line radiosurgery for small-cell and non-small cell lung cancer brain metastases (CROSS-FIRE).

INTRODUCTION: Historical reservations regarding stereotactic radiosurgery (SRS) for small-cell lung cancer (SCLC) brain metastases include concerns for short-interval and diffuse central nervous system (CNS) progression, poor prognoses, and increased neurological mortality specific to SCLC histology. We compared SRS outcomes for SCLC and non-small cell lung cancer (NSCLC) where SRS is well established. METHODS: Multicenter first-line SRS outcomes for SCLC and NSCLC from 2000 to 2022 were retrospectively collected (n = 892 SCLC, n = 4785 NSCLC). Data from the prospective Japanese Leksell Gamma Knife Society (JLGK0901) clinical trial of first-line SRS were analyzed as a comparison cohort (n = 98 SCLC, n = 814 NSCLC). Overall survival (OS) and CNS progression were analyzed using Cox proportional hazard and Fine-Gray models, respectively, with multivariable adjustment for cofactors including age, sex, performance status, year, extracranial disease status, and brain metastasis number and volume. Mutation-stratified analyses were performed in propensity score-matched retrospective cohorts of epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) positive NSCLC, mutation-negative NSCLC, and SCLC. RESULTS: OS was superior for patients with NSCLC compared to SCLC in the retrospective dataset (median OS = 10.5 vs 8.6 months; P < .001) and in the JLGK0901 dataset. Hazard estimates for first CNS progression favoring NSCLC were similar in both datasets but reached statistical significance in the retrospective dataset only (multivariable hazard ratio = 0.82, 95% confidence interval = 0.73 to 0.92, P = .001). In the propensity score-matched cohorts, there were continued OS advantages for NSCLC patients (median OS = 23.7 [EGFR and ALK positive NSCLC] vs 13.6 [mutation-negative NSCLC] vs 10.4 months [SCLC], pairwise P values < 0.001), but no statistically significant differences in CNS progression were observed in the matched cohorts. Neurological mortality and number of lesions at CNS progression were similar for NSCLC and SCLC patients. Leptomeningeal progression was increased in patients with NSCLC compared to SCLC in the retrospective dataset only (multivariable hazard ratio = 1.61, 95% confidence interval = 1.14 to 2.26, P = .007). CONCLUSIONS: After SRS, SCLC histology was associated with shorter OS compared to NSCLC. CNS progression occurred earlier in SCLC patients overall but was similar in patients matched on baseline factors. SCLC was not associated with increased neurological mortality, number of lesions at CNS progression, or leptomeningeal progression compared to NSCLC. These findings may better inform clinical expectations and individualized decision making regarding SRS for SCLC patients.

Multicenter analysis of stereotactic radiosurgery for multiple brain metastases from EGFR and ALK driven non-small cell lung cancer.

INTRODUCTION: Patients with brain metastases (BrMs) arising from EGFR and ALK driven non-small cell lung cancer (NSCLC) have favorable prognoses and evolving treatment options. We evaluated multicenter outcomes for stereotactic radiosurgery (SRS) to multiple (≥4) BrMs, where randomized data remain limited. METHODS: Data were collected retrospectively from 5 academic centers on EGFR and ALK NSCLC who received SRS to ≥4 BrMs with their first SRS treatment between 2008 and 2018. Analyzed endpoints included overall survival (OS), freedom from CNS progression (FFCNSP), and freedom from whole-brain radiotherapy (FFWBRT). RESULTS: Eighty-nine patients (50 EGFR, 39 ALK) received a total of 159 SRS treatments to 1,080 BrMs, with a median follow up of 51.3 months. The median number of BrMs treated with SRS treatment-1 was 6 (range 4-26) and median for all treatments was 9 (range 4-47). Sixteen patients (18 %) had received WBRT prior to SRS treatment-1. The median OS was 24.2, 21.2, and 33.2 months for all patients, EGFR, and ALK subsets, respectively. After multivariable adjustment, only receipt of a next-generation tyrosine kinase inhibitor was associated with OS (HR 0.40, p = 0.005). No differences in OS were observed based on number of BrMs treated. The median FFCNSP was 9.4, 11.6, and 7.5 months, for all patients, EGFR, and ALK subsets, respectively. After multivariable adjustment, the number of BrMs (continuous) treated during treatment-1 was the only negative prognostic factor associated with FFCNSP (HR 1.071, p = 0.045). The 5-year FFWBRT was 73.6 %. CONCLUSIONS: This multicenter analysis over a >10-year period demonstrated favorable OS, FFCNSP, and FFWBRT, in patients with EGFR and ALK driven NSCLC receiving SRS to ≥4 BrMs. These data support SRS as an option in the upfront and salvage setting for higher burden CNS disease in this population.

Multimodality Management of EBV-Associated Nasopharyngeal Carcinoma.

Nasopharyngeal carcinoma (NPC) is a rare cancer of the nasopharyngeal mucosa with a specific geographic predisposition. NPC is often associated with Epstein-Barr Virus (EBV) infection and as a result contains many characteristic biomarkers. Treatment of locally-contained NPC is generally achieved through use of radiotherapy (RT), as part of a multimodality treatment regimen. Induction chemotherapy followed by concurrent RT and platinum-based chemotherapy regimen has emerged as the definitive treatment of choice for locoregionally-advanced NPC. Recently, immunotherapy is finding a role in the treatment of recurrent or metastatic NPC. Immune checkpoint blockade therapies targeted against the programmed death-1 (PD-1) receptor have demonstrated efficacy in early phase clinical trials, with ongoing phase III trials in effect. Biomarkers for treatment efficacy remain an ongoing area of investigation, with important prognostic implications on the horizon.

Radiation Therapy for Benign Disease: Arteriovenous Malformations, Desmoid Tumor, Dupuytren Contracture, Graves Ophthalmopathy, Gynecomastia, Heterotopic Ossification, Histiocytosis.

Although the use of ionizing radiation in malignant conditions has been well established, its application in benign conditions has not been fully accepted and has been inadequately recognized by health care providers outside of radiation therapy. Most frequently, radiation therapy in these benign conditions is used along with other treatment modalities, such as surgery, in instances where the condition causes significant disability or could even lead to death. Radiation therapy can be helpful for inflammatory/proliferative disorders. This article discusses the current use of radiation therapy in some of the more common benign conditions.

Radiation Therapy for Benign Disease: Keloids, Macular Degeneration, Orbital Pseudotumor, Pterygium, Peyronie Disease, Trigeminal Neuralgia.

Although the use of ionizing radiation on malignant conditions has been well established, its application on benign conditions has not been fully accepted and has been inadequately recognized by health care providers outside of radiation therapy. Most frequently, radiation therapy in these benign conditions is used along with other treatment modalities, such as surgery, when the condition causes significant disability or could even lead to death. Radiation therapy can be helpful for inflammatory/proliferative disorders. This article discusses the present use of radiation therapy for some of the most common benign conditions.

Dramatic Resolution of an Unresectable Giant Basal Cell Carcinoma Treated with Intensity-Modulated Radiation Therapy (IMRT) - A Case Report.

A 59-year-old man presented with an unresectable bulky giant basal cell carcinoma on his upper back. A trial of chemotherapy did not help relieve his symptoms or reduce the tumor. He was referred for and received definitive radiation therapy via IMRT with dramatic regression. The patient had been unable to lie on his back for many years but currently can sleep comfortably on his back without pain, which has dramatically improved his quality of life.

Epithelial Na⁺ sodium channels in magnocellular cells of the rat supraoptic and paraventricular nuclei.

The epithelial Na⁺ channels (ENaCs) are present in kidney and contribute to Na⁺ and water homeostasis. All three ENaC subunits (α, ß, and γ) were demonstrated in the cardiovascular regulatory centers of the rat brain, including the magnocellular neurons (MNCs) in the supraoptic nucleus (SON) and the paraventricular nucleus (PVN). However, the functional significance of ENaCs in vasopressin (VP) and oxytocin (OT) synthesizing MNCs is completely unknown. In this study, we show with immunocytochemical double-labeling that the α-ENaC is colocalized with either VP or OT in MNCs in the SON and PVN. In addition, parvocellular neurons in the dorsal, ventrolateral, and posterior subregions of the PVN (not immunoreactive to VP or OT) are also immunoreactive for α-ENaC. In contrast, immunoreactivity to ß- and γ-ENaC is colocalized with VP alone within the MNCs. Furthermore, immunoreactivity for a known target for ENaC expression, the mineralcorticoid receptor (MR), is colocalized with both VP and OT in MNCs. Using single-cell RT-PCR, we detected mRNA for all three ENaC subunits and MR in cDNA libraries derived from single MNCs. In whole cell voltage clamp recordings, application of the ENaC blocker benzamil reversibly reduced a steady-state inward current and decreased cell membrane conductance approximately twofold. Finally, benzamil caused membrane hyperpolarization in a majority of VP and about one-half of OT neurons in both spontaneously firing and quiet cells. These results strongly suggest the presence of functional ENaCs that may affect the firing patterns of MNCs, which ultimately control the secretion of VP and OT.