[en] Recent advances in deep neural networks (DNNs) have unlocked opportunities for their application for automatic image segmentation. We have evaluated a DNN-based algorithm for automatic segmentation of the prostate gland on a large cohort of patient images.

[en] Purpose: To evaluate the clinical potential of deformable image registration (DIR)-based automatic propagation of physician-drawn contours from a planning CT to midtreatment CT images for head and neck (H and N) adaptive radiotherapy. Methods: Ten H and N patients, each with a planning CT (CT1) and a subsequent CT (CT2) taken approximately 3–4 week into treatment, were considered retrospectively. Clinically relevant organs and targets were manually delineated by a radiation oncologist on both sets of images. Four commercial DIR algorithms, two B-spline-based and two Demons-based, were used to deform CT1 and the relevant contour sets onto corresponding CT2 images. Agreement of the propagated contours with manually drawn contours on CT2 was visually rated by four radiation oncologists in a scale from 1 to 5, the volume overlap was quantified using Dice coefficients, and a distance analysis was done using center of mass (CoM) displacements and Hausdorff distances (HDs). Performance of these four commercial algorithms was validated using a parameter-optimized Elastix DIR algorithm. Results: All algorithms attained Dice coefficients of >0.85 for organs with clear boundaries and those with volumes >9 cm³. Organs with volumes <3 cm³ and/or those with poorly defined boundaries showed Dice coefficients of ∼0.5–0.6. For the propagation of small organs (<3 cm³), the B-spline-based algorithms showed higher mean Dice values (Dice = 0.60) than the Demons-based algorithms (Dice = 0.54). For the gross and planning target volumes, the respective mean Dice coefficients were 0.8 and 0.9. There was no statistically significant difference in the Dice coefficients, CoM, or HD among investigated DIR algorithms. The mean radiation oncologist visual scores of the four algorithms ranged from 3.2 to 3.8, which indicated that the quality of transferred contours was “clinically acceptable with minor modification or major modification in a small number of contours.” Conclusions: Use of DIR-based contour propagation in the routine clinical setting is expected to increase the efficiency of H and N replanning, reducing the amount of time needed for manual target and organ delineations

[en] Purpose: To determine the feasibility, safety, and efficacy of stereotactic body radiation therapy (SBRT), also known as radiosurgery, in patients with head-and-neck cancers. Methods and Materials: Patients with pathologically proven malignant lesions in the head-and-neck region were treated using single-dose SBRT (S-SBRT) or fractionated SBRT (F-SBRT). Radiation doses were either single-fraction 13-18 Gy for S-SBRT or 36-48 Gy in five to eight fractions for F-SBRT. Response evaluation was based on clinical examinations and computed tomography/magnetic resonance imaging scans. Pre- and post-SBRT tumor dimensions were measured in three axes, and tumor volumes were calculated. Response evaluation also was performed using World Health Organization criteria. Results: Fifty-five lesions were treated in 44 patients (25 men, 19 women). There were three groups of patients: those with primary (n = 10), recurrent (n = 21), and metastatic tumors (n = 13). The predominant histologic type was squamous cell carcinoma (n = 33). The majority of lesions were treated using F-SBRT (n = 37). Based on radiographic and clinical assessment, a 77% (complete + partial response) response rate was noted. Percentage of reduction in tumor volume was 52% ± 38% based on follow-up scans in 24 patients. Tumor control rates at 1 year were 83.3% and 60.6% in the primary and recurrent groups, respectively. Median overall survival was 28.7, 6.7, and 5.6 months for the primary, recurrent, and metastatic groups, respectively. Radiation Therapy Oncology Group Grade 1-2 mucositis was noted in all patients treated for oropharyngeal or laryngeal lesions. Conclusions: The SBRT in single or fractionated doses offers a viable treatment option for selected patients with primary, recurrent, and metastatic head-and-neck cancers with functional preservation.

[en] Purpose: To develop standard (SPCA) and regularized (RPCA) principal component analysis models of anatomical changes from daily cone beam CTs (CBCTs) of head and neck (H&N) patients and assess their potential use in adaptive radiation therapy, and for extracting quantitative information for treatment response assessment. Methods: Planning CT images of ten H&N patients were artificially deformed to create “digital phantom” images, which modeled systematic anatomical changes during radiation therapy. Artificial deformations closely mirrored patients’ actual deformations and were interpolated to generate 35 synthetic CBCTs, representing evolving anatomy over 35 fractions. Deformation vector fields (DVFs) were acquired between pCT and synthetic CBCTs (i.e., digital phantoms) and between pCT and clinical CBCTs. Patient-specific SPCA and RPCA models were built from these synthetic and clinical DVF sets. EigenDVFs (EDVFs) having the largest eigenvalues were hypothesized to capture the major anatomical deformations during treatment. Results: Principal component analysis (PCA) models achieve variable results, depending on the size and location of anatomical change. Random changes prevent or degrade PCA’s ability to detect underlying systematic change. RPCA is able to detect smaller systematic changes against the background of random fraction-to-fraction changes and is therefore more successful than SPCA at capturing systematic changes early in treatment. SPCA models were less successful at modeling systematic changes in clinical patient images, which contain a wider range of random motion than synthetic CBCTs, while the regularized approach was able to extract major modes of motion. Conclusions: Leading EDVFs from the both PCA approaches have the potential to capture systematic anatomical change during H&N radiotherapy when systematic changes are large enough with respect to random fraction-to-fraction changes. In all cases the RPCA approach appears to be more reliable at capturing systematic changes, enabling dosimetric consequences to be projected once trends are established early in a treatment course, or based on population models.

[en] Raman spectroscopy is an optical technique capable of identifying chemical constituents of a sample by their unique set of molecular vibrations. Research on the applicability of Raman spectroscopy in the differentiation of cancerous versus normal tissues has been ongoing for many years, and has yielded successful results in the context of prostate, breast, brain, skin, and head and neck cancers as well as pediatric tumors. Recently, much effort has been invested on developing noninvasive “Raman” probes to provide real-time diagnosis of potentially cancerous tumors. In this regard, it is feasible that the Raman technique might one day be used to provide rapid, minimally invasive real-time diagnosis of tumors in patients. Raman spectroscopy is relatively new to the field of radiation therapy. Recent work involving cell lines has shown that the Raman technique is able to identify proteins and other markers affected by radiation therapy. Although this work is preliminary, one could ask whether or not the Raman technique might be used to identify molecular markers that predict radiation response. This paper provides a brief review of Raman spectroscopic investigations in cancer detection, benefits and limitations of this method, advances in instrument development, and also preliminary studies related to the application of this technology in radiation therapy response assessment

[en] Purpose: To determine our institutional vertebral compression fracture (VCF) rate after spine stereotactic radiosurgery (SRS) and determine contributory factors. Methods and Materials: Retrospective analysis from 2001 to 2013 at a single institution was performed. With institutional review board approval, electronic medical records of 1905 vertebral bodies from 791 patients who were treated with SRS for the management of primary or metastatic spinal lesions were reviewed. A total of 448 patients (1070 vertebral bodies) with adequate follow-up imaging studies available were analyzed. Doses ranging from 10 Gy in 1 fraction to 60 Gy in 5 fractions were delivered. Computed tomography and magnetic resonance imaging were used to evaluate the primary endpoints of this study: development of a new VCF, progression of an existing VCF, and requirement of stabilization surgery after SRS. Results: A total of 127 VCFs (11.9%; 95% confidence interval [CI] 9.5%-14.2%) in 97 patients were potentially SRS induced: 46 (36%) were de novo, 44 (35%) VCFs progressed, and 37 (29%) required stabilization surgery after SRS. Our rate for radiologic VCF development/progression (excluding patients who underwent surgery) was 8.4%. Upon further exclusion of patients with hematologic malignancies the VCF rate was 7.6%. In the univariate analyses, females (hazard ratio [HR] 1.54, 95% CI 1.01-2.33, P=.04), prior VCF (HR 1.99, 95% CI 1.30-3.06, P=.001), primary hematologic malignancies (HR 2.68, 95% CI 1.68-4.28, P<.001), thoracic spine lesions (HR 1.46, 95% CI 1.02-2.10, P=.02), and lytic lesions had a significantly increased risk for VCF after SRS. On multivariate analyses, prior VCF and lesion type remained contributory. Conclusions: Single-fraction SRS doses of 16 to 18 Gy to the spine seem to be associated with a low rate of VCFs. To the best of our knowledge, this is the largest reported experience analyzing SRS-induced VCFs, with one of the lowest event rates reported.

[en] Purpose: To analyze the prospectively collected health-related quality-of-life (HRQOL) data from patients enrolled in two Radiation Therapy Oncology Group randomized Phase III head and neck cancer trials (90-03 and 91-11) to assess their value as an independent prognostic factor for locoregional control (LRC) and/or overall survival (OS). Methods and Materials: HRQOL questionnaires, using a validated instrument, the Functional Assessment of Cancer Therapy-Head and Neck (FACT-H and N), version 2, were completed by patients before the start of treatment. OS and LRC were the outcome measures analyzed using a multivariate Cox proportional hazard model. Results: Baseline FACT-H and N data were available for 1,093 patients and missing for 417 patients. No significant difference in outcome was found between the patients with and without baseline FACT-H and N data (p = 0.58). The median follow-up time was 27.2 months for all patients and 49 months for surviving patients. Multivariate analyses were performed for both OS and LRC. Beyond tumor and nodal stage, Karnofsky performance status, primary site, cigarette use, use of concurrent chemotherapy, and altered fractionation schedules, the FACT-H and N score was independently predictive of LRC (but not OS), with p = 0.0038. The functional well-being component of the FACT-H and N predicted most significantly for LRC (p = 0.0004). Conclusions: This study represents, to our knowledge, the largest analysis of HRQOL as a prognostic factor in locally advanced head and neck cancer patients. The results of this study have demonstrated the importance of baseline HRQOL as a significant and independent predictor of LRC in patients with locally advanced head and neck cancer

[en] Purpose: To assess the safety and efficacy of combining oncolytic adenovirus-mediated cytotoxic gene therapy (OAMCGT) with intensity modulated radiation therapy (IMRT) in intermediate-risk prostate cancer. Methods and Materials: Forty-four men with intermediate-risk prostate cancer were randomly assigned to receive either OAMCGT plus IMRT (arm 1; n=21) or IMRT only (arm 2; n=23). The primary phase 2 endpoint was acute (≤90 days) toxicity. Secondary endpoints included quality of life (QOL), prostate biopsy (12-core) positivity at 2 years, freedom from biochemical/clinical failure (FFF), freedom from metastases, and survival. Results: Men in arm 1 exhibited a greater incidence of low-grade influenza-like symptoms, transaminitis, neutropenia, and thrombocytopenia than men in arm 2. There were no significant differences in gastrointestinal or genitourinary events or QOL between the 2 arms. Two-year prostate biopsies were obtained from 37 men (84%). Thirty-three percent of men in arm 1 were biopsy-positive versus 58% in arm 2, representing a 42% relative reduction in biopsy positivity in the investigational arm (P=.13). There was a 60% relative reduction in biopsy positivity in the investigational arm in men with <50% positive biopsy cores at baseline (P=.07). To date, 1 patient in each arm exhibited biochemical failure (arm 1, 4.8%; arm 2, 4.3%). No patient developed hormone-refractory or metastatic disease, and none has died from prostate cancer. Conclusions: Combining OAMCGT with IMRT does not exacerbate the most common side effects of prostate radiation therapy and suggests a clinically meaningful reduction in positive biopsy results at 2 years in men with intermediate-risk prostate cancer