Tumor Subvolumes Research Articles

Radiation treatment planning study to investigate feasibility of delivering Immunotherapy in Combination with Ablative Radiosurgery to Ultra-High DoSes (ICARUS).

PurposeImmune checkpoint inhibitors improve survival in metastatic diseases for some cancers. Multisite SBRT with pembrolizumab (SBRT + Pembro) was shown to be safe with promising local control using biologically effective doses (BEDs) = 95–120 Gy. Increased BED may improve response rate; however, SBRT doses are limited by surrounding organs at risk (OARs). The purpose of this work was to develop and validate methods for safe delivery of ultra‐high doses of radiation (BED10 > 300) to be used in future clinical trials.Methods and MaterialsThe radiation plans from 15 patients enrolled on a phase I trial of SBRT + pembro were reanalyzed. Metastatic disease sites included liver (8/15), inguinal region (1/15), pelvis (2/15), lung (1/15), abdomen (1/15), spleen (1/15), and groin (1/15). Gross tumor volumes (GTVs) ranged from 80 to 708 cc. Following the same methodology used in the Phase I trial on which these patients were treated, GTVs > 65 cc were contracted to a 65 cc subvolume (SubGTV) resulting in only a portion of the GTV receiving prescription dose. Volumetric modulated arc therapy (VMAT) was used to plan treatments BED10 = 360 Gy. Plans utilizing both 6FFF and 10FFF beams were compared to clinical plans delivering BED10 = 112.50 Gy. The target primary goal was V100% > 95% with a secondary goal of V70% > 99% and OAR objectives per the trial. To demonstrate feasibility, plans were delivered to a diode array phantom and evaluated for fidelity using gamma analysis.ResultsAll 30 plans met the secondary coverage goal and satisfied all OAR constraints. The primary goal was achieved in 12/15 of the 6FFF plans and 13/15 of the 10FFF plans. Average gamma analysis passing rate using criteria of 3% dose difference and 3, 2, and 1 mm were 99.1 ± 1.0%, 98.5 ± 1.6%, and 95.1 ± 3.8%, respectively.ConclusionNovel VMAT planning approaches with clinical treatment planning software and linear accelerators prove capable of delivering radiation doses in excess of 360 Gy BED10 to tumor subvolumes, while maintaining safe OAR doses.

PD-0429: CT radiomics differentiates levels of radiocurability in tumor subvolumes in head and neck cancer

PD-0429: CT radiomics differentiates levels of radiocurability in tumor subvolumes in head and neck cancer

PO-1695: The utility of multiparametric MRI to characterize hypoxic tumor subvolumes compared to FMISO PET/CT

PO-1695: The utility of multiparametric MRI to characterize hypoxic tumor subvolumes compared to FMISO PET/CT

Marginless Radiotherapy (MRT): Leveraging The Abscopal Effect Toward The Treatment Of Metastatic Castration Resistant Prostate Cancer

Marginless Radiotherapy (MRT) is a new radiotherapy approach whereby only the gross tumor volume (GTV) or sub-volume (sv) is irradiated. We sought to investigate the potential of MRT in combination with in-situ delivered anti-CD40 to generate a robust abscopal effect in metastatic castration resistant prostate cancer. Using an aggressive castration resistant prostrate adenocarcinoma cell line derived from TRAMP-C1, two contralateral tumors were implanted in each mouse either subcutaneously (SQ) in two flanks or one SQ and one orthotopic to anterior prostate lobe. In both cases one SQ tumor was chosen to be treated. The mice were randomly assigned to different cohorts (n = 5-8) for image-guided radiotherapy (IGRT) of a single dose of 2, 5, 8, or 10 Gy using Small Animal Radiation Research Platform (SARRP). Following the IGRT, each cohort was treated with 20 μg of intratumoral mouse antiCD40 antibody loaded in a smart radiotherapy biomaterial (srbAntiCD40) for sustain release. These cohorts were then compared with a control group with no treatment. Tumor volume of both treated (treated site) and untreated tumors (abscopal site) were then measured and survival percent was calculated to determine the most potent RT dose which was then used for the PTV vs. GTV/sv of IGRT study. PTV of IGRT was delivered with 10x10 mm2 field size collimators whereas 1x1, 3x3, or 5x5 mm2 field size collimators were used for GTV/sv. Statistical analyses for tumor volume and Kaplan–Meier survival assay were then analyzed. A single fraction IGRT dose of 5 Gy to the PTV with srbAntiCD40 shows the highest survival fraction (****) compare with other radiation doses. Further analysis of 5 Gy of IGRT in PTV or GTV/sv along with srbAntiCD40 (intratumoral) demonstrates non-significant difference of the survival rate with PTV (80%**) compare to GTV/sv in SQ model (100%***) or orthotopic model (80%***). Similarly, fold increase of tumor volume (from day of treatment) in treated sides were 91 (+/- 11.5), 11* (+/- 6.5), 6* (+/-2.5), and 6* (+/- 2.5) respectively for Control, RT-PTV, RT-GTVsv (SQ), and RT-GTVsv (orthotopic) cohorts. Untreated, abscopal side tumors also shown similar effects (60 (+/- 17.4), 13** (+/- 2.4), 11**(+/-2.1), and 6** (+/- 1.3) respectively. Complete regression of both of the tumors were shown in 20% of mice in both SQ and orthotopic RT-GTVsv groups. Significant high infiltration of CD8+ cytotoxic T cells were also observed in both treated and untreated tumors of both PTV and GTVsv cohorts, 10 days after the treatment given. This study proffers a new MRT strategy for treating aggressive metastatic prostate cancer, leveraging the abscopal effect more efficaciously by treating only a GTV/sv which further reduces the radiation induced toxicity. This strategy may spare peripheral APCs cells, which are crucial in driving the immune-mediated abscopal effect.

SBRT-Partial Tumor Irradiation for The Treatment of Unresectable Bulky NSCLC

to report on the improved therapeutic ratio by partial tumor irradiation exploiting the bystander (BE) and abscopal effects (AE) for the treatment of bulky NSCLC. Unlike conventional (whole tumor + CTV + PTV) RT, in order to generate predominant immune-mediated tumor cell killing by radiation, we developed SBRT-PArtial Tumor irradiation targeting HYpoxic segment (SBRT-PATHY) while sparing and preserving the functionality of peritumoral immune microenvironment (PIM) and the regional lymphocytes necessary to mediate BE/AE. Our preclinical studies showed that the hypoxic tumor sub-volume if selectively irradiated was the “most-abscopal one” with highest inductive abscopal potential compared with normoxic segment. Our hypothesis is that SBRT-PATHY can improve the therapeutic ratio of unresectable bulky NSCLC by exploiting BE and AE. The primary endpoint was bulky tumor control rate. The secondary endpoints included AE rate, survival and toxicity. 20 patients with inoperable stage IIIB/IV bulky NSCLC considered unsuitable for conventional RT-chemotherapy due to comorbidities or tumor site and size were included. Combination of CT and 18F-FDG-PET was used per institutional protocol to define the hypovascularized (contrast-hypo-enhanced) and hypometabolic (SUV max < 3) tumor segment which was irradiated with 10Gy × 3 to the 70%-isodose. PIM was contoured as 1cm large peritumoral tissue that was spared from radiation (applied dose constraints: Dmin<1Gy/fraction, Dmax<5Gy/fraction, Dmean<3Gy/fraction). Immunohistochemistry and gene expression analysis were performed on the available tissue samples of 3 patients submitted to surgery following SBRT-PATHY in order to extract responding partially irradiated and unirradiated-abscopal tumor lesions. Median follow up was 12.9 months (range: 4-27). Mean bulky diameter was 8.4 cm (range: 6-14.5). On an average, the irradiated volume corresponded to 32% of the total tumor volume. The bulky tumor control rate was 95% (average bulky shrinkage 71%). AE was observed in 46% of patients. Overall, cancer-specific and progression-free survival were 55%, 75% and 65%, respectively. Multi-variate analysis for cancer specific survival was significant for SBRT-PATHY independent of age, sex, performance status, histology, stage, bulky site and volume (p < 0.001). 6 patients experienced fatigue grade 1. An effective antitumor immune response was activated in radiation-spared PIM leading to cell death of the partially irradiated tumors. CD20+ B-lymphocytes, CD3+/CD8+ T-lymphocytes densely infiltrated PIM and the border of the partially irradiated tumors. Apoptotic abscopal tumor sites showed a strong expression of the cell death-inducing cytokines, TGF-beta and FAS ligand and to a lesser degree of TRAIL and TNF-alpha. SBRT-PATHY for unresectable bulky NSCLC showed inspiring results that may potentially enhance the radiotherapy therapeutic ratio. A larger prospective trial is ongoing.

Comprehensive Analysis of Tumour Sub-Volumes for Radiomic Risk Modelling in Locally Advanced HNSCC.

Simple SummaryRadiomic risk models are usually based on imaging features, which are extracted from the entire gross tumour volume (GTV). This approach does not explicitly consider the complex biological structure of the tumours. Therefore, in this retrospective study, we investigated the prognostic value of radiomic analyses based on different tumour sub-volumes using computed tomography imaging of patients with locally advanced head and neck squamous cell carcinoma who were treated with primary radio-chemotherapy. The GTV was cropped by different margins to define the rim and corresponding core sub-volumes of the tumour. Furthermore, the best performing tumour rim sub-volume was extended into surrounding tissue with different margins. As a result, the models based on the 5 mm tumour rim and on the 3 mm extended rim sub-volume showed an improved performance compared to models based on the corresponding tumour core. This indicates that the consideration of tumour sub-volumes may help to improve radiomic risk models.Imaging features for radiomic analyses are commonly calculated from the entire gross tumour volume (GTV). However, tumours are biologically complex and the consideration of different tumour regions in radiomic models may lead to an improved outcome prediction. Therefore, we investigated the prognostic value of radiomic analyses based on different tumour sub-volumes using computed tomography imaging of patients with locally advanced head and neck squamous cell carcinoma. The GTV was cropped by different margins to define the rim and the corresponding core sub-volumes of the tumour. Subsequently, the best performing tumour rim sub-volume was extended into surrounding tissue with different margins. Radiomic risk models were developed and validated using a retrospective cohort consisting of 291 patients in one of the six Partner Sites of the German Cancer Consortium Radiation Oncology Group treated between 2005 and 2013. The validation concordance index (C-index) averaged over all applied learning algorithms and feature selection methods using the GTV achieved a moderate prognostic performance for loco-regional tumour control (C-index: 0.61 ± 0.04 (mean ± std)). The models based on the 5 mm tumour rim and on the 3 mm extended rim sub-volume showed higher median performances (C-index: 0.65 ± 0.02 and 0.64 ± 0.05, respectively), while models based on the corresponding tumour core volumes performed less (C-index: 0.59 ± 0.01). The difference in C-index between the 5 mm tumour rim and the corresponding core volume showed a statistical trend (p = 0.10). After additional prospective validation, the consideration of tumour sub-volumes may be a promising way to improve prognostic radiomic risk models.

Manual and semi-automated delineation of locally advanced rectal cancer subvolumes with diffusion-weighted MRI.

To evaluate interobserver agreement for T2 weighted (T2W) and diffusion-weighted MRI (DW-MRI) contours of locally advanced rectal cancer (LARC); and to evaluate manual and semi-automated delineations of restricted diffusion tumour subvolumes. 20 cases of LARC were reviewed by 2 radiation oncologists and 2 radiologists. Contours of gross tumour volume (GTV) on T2W, DW-MRI and co-registered T2W/DW-MRI were independently delineated and compared using Dice Similarity Coefficient (DSC), mean distance to agreement (MDA) and other metrics of interobserver agreement. Restricted diffusion subvolumes within GTVs were manually delineated and compared to semi-automatically generated contours corresponding to intratumoral apparent diffusion coefficient (ADC) centile values. Observers were able to delineate subvolumes of restricted diffusion with moderate agreement (DSC 0.666, MDA 1.92 mm). Semi-automated segmentation based on the 40th centile intratumoral ADC value demonstrated moderate average agreement with consensus delineations (DSC 0.581, MDA 2.44 mm), with errors noted in image registration and luminal variation between acquisitions. A small validation set of four cases with optimised planning MRI demonstrated improvement (DSC 0.669, MDA 1.91 mm). Contours based on co-registered T2W and DW-MRI could be used for delineation of biologically relevant tumour subvolumes. Semi-automated delineation based on patient-specific intratumoral ADC thresholds may standardise subvolume delineation if registration between acquisitions is sufficiently accurate. This is the first study to evaluate the feasibility of semi-automated diffusion-based subvolume delineation in LARC. This approach could be applied to dose escalation or 'dose painting' protocols to improve delineation reproducibility.

Abstract 2773: Increased tumor control boosting hypoxic regions identified with EPR pO2 imaging in 2 tumor types

Abstract Introduction: It has been known for over a century that all living systems develop radiation resistance under hypoxic conditions - low pO2. But giving extra radiation - dose painting - or boosting hypoxic tumors regions has not shown enhanced control in mammalian tumors. Pulse EPR quantitative pO2 values can be tomographically imaged as a magnetic resonance measurement with fixed stepped gradients. If there are distinct hypoxic regions, their images can then be extra radiation dose boosts. An isocentric animal irradiation system allowed conformal boosts to tiny resistant tumor subvolumes with 3D printed tungsten loaded plastic blocks. A randomized trial compared radiation boosts to hypoxic tumor vs well oxygenated tumor to determine hypoxic boost control advantage. Methods: Carcinomas (MCa4) and fibrosarcomas (FSa) in mouse legs with mean volume of 350 μl were treated with an XRAD225Cx animal irradiator to a dose that controlled 15% of tumors (TCD15), determined in separate experiments. Pulse electron paramagnetic resonance (EPR) using Spin Lattice Relaxation (SLR) and a soluble trityl spin probe has been shown to quantitatively image solution and tumor pO2. This and boost treatments were given after EPR pO2 imaging, reconstruction, registration with tumor defining T2 MRI, determination of all tumor hypoxic regions (pO2 ≤10 torr) and 3D printed tungsten loaded plastic aperture fabrication to radiate either well oxygenated or hypoxic tumor regions. Mice were followed for either 180 days (MCa4) or 90 days (FSa) sufficient for 99% of local failures to occur. Treatment of 3D hypoxic volumes was also analyzed for control dependence on surface to volume ratio (SVR). Results: Tumor control was doubled by radiating hypoxic tumor relative to well oxygenated tumor in 48 MCa4 tumors (p=0.013) and 54 FSa tumors (p=0.04). This was the first demonstration in mammalian tumors of an advantage to targeting hypoxic regions in tumors. It provides the first evidence of image based dose painting within the volume of the tumor. For MCa4 tumors, SVR values below median had significant increased survival (p=0.003) while for FSa tumors below median SVR tumors trended to increased survival (p=0.3). Clinical Relevance: Modern radiation delivery used intensity modulated radiation therapy (IMRT) for dose gradients outside the tumor volume to reduce doses to critical structures. Dose within the tumor have strict inhomogeneity limits. Allowance of higher dose variation within the tumor volume allows “hot spots” to resistant tumor volumes, sharper gradients outside the volume to protect critical structures. This will enhance the therapeutic ratio. Citation Format: Howard Halpern, Matthew Maggio, Eugene Barth, Darwin Bodero, Richard C. Miller, Charles A. Pelizzari, Martyna Krzykawska-Serda, Subramanian V. Sundramoorthy, Bulent Aydogan, Ralph R. Weichselbaum, Victor M. Tormyshev, Inna Gertsenshteyn, Boris Epel. Increased tumor control boosting hypoxic regions identified with EPR pO2 imaging in 2 tumor types [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2773.

Conformal Avoidance of Normal Organs at Risk by Perfusion-Modulated Dose Sculpting in Tumor Single-Dose Radiation Therapy

Although 24 Gy single-dose radiation therapy (SDRT) renders >90% 5-year local relapse-free survival in human solid tumor lesions, SDRT delivery is not feasible in ∼50% of oligometastatic lesions owing to interference by dose/volume constraints of a serial organ at risk (OAR). Conformal OAR avoidance is based on a hypothetical model positing that the recently described SDRT biology specifically permits volumetric subdivision of the SDRT dose, such that high-intensity vascular drivers of SDRT lethality, generated within a major tumor subvolume exposed to a high 24 Gy dose (high-dose planning target volume [PTVHD]), would equilibrate SDRT signaling intensity throughout the tumor interstitial space, rendering bystander radiosensitization of a minor subvolume (perfusion-modulated dose sculpting PTV [PTVPMDS]), dose-sculpted to meet a serial OAR dose/volume constraint. An engineered PTVPMDS may thus yield tumor ablation despite PMDS dose reduction and conformally avoiding OAR exposure to a toxic dose. Dose fall-off within the PTVPMDS penumbra of oligometastatic lesions was planned and delivered by intensity modulated inverse dose painting. SDRT- and SDRT-PMDS-treated lesions were followed with periodic positron emission tomography/computed tomography imaging to assess local tumor control. Cumulative baseline 5-year local relapse rates of oligometastases treated with 24 Gy SDRT alone (8% relapses, n = 292) were similar in moderate PTVPMDS dose-sculpted (23-18 Gy, n = 76, 11% relapses, P = .36) and extreme dose-sculpted (<18 Gy, n = 61, 14% relapses, P = .29) lesions, provided the major 24 Gy PTVHD constituted ≥60% of the total PTV. In contrast, 28% of local relapses occurred in 26 extreme dose-sculpted PTVPMDS lesions when PTVHD constituted <60% of the total PTV (P = .004), suggesting a threshold for the PTVPMDS bystander effect. The study provides compelling clinical support for the bystander radiosensitization hypothesis, rendering local cure of tumor lesions despite a ≥25% PTVPMDS dose reduction of the 24 Gy PTVHD dose, adapted to conformally meet OAR dose/volume constraints. The SDRT-PMDS approach thus provides a therapeutic resolution to otherwise radioablation-intractable oligometastatic disease.

The utility of multiparametric MRI to characterize hypoxic tumor subvolumes in comparison to FMISO PET/CT. Consequences for diagnosis and chemoradiation treatment planning in head and neck cancer

Background and PurposeHypoxia is an essential metabolic marker that determines chemo- and radiation resistance in head-and-neck squamous cell carcinoma (HNSCC) patients. Our exploratory analysis aimed to identify multiparametric MRI (mpMRI) parameters linked to hypoxia that might be used as surrogate for [18F]FMISO-PET in diagnosis and chemoradiation treatment (CRT) of HNSCC. Materials and Methods21 patients undergoing definitive CRT for HNSCC were prospectively imaged with serial [18F]FMISO-PET and 3 Tesla mpMRI for T1- and T2-weighted and dynamic contrast-enhanced perfusion and diffusion-weighted measurements (ktrans, ve, kep, ADC) in weeks 0, 2 and 5 and FDG-PET in week 0. [18F]FMISO-PET-derived hypoxic subvolumes (HSV) and complementary non-hypoxic subvolumes (nonHSV) were created for tumor and lymph nodes and projected on the mpMRI scans after PET/MRI co-registration. MpMRI and [18F]FMISO-PET parameters within HSVs and nonHSVs were statistically compared. ResultsFMISO-PET-based HSVs of the primary tumors on MRI were characterized by lower ADC at all time points (p = 0.012 at baseline; p = 0.015 in week 2) and reduced interstitial space volume fraction ve and perfusion ktrans at baseline (p = 0.006, p = 0.047) compared to nonHSVs. Hypoxic lymph nodes were characterized by significantly lower ADC values at baseline (p = 0.039), but not at later time points and a reduction in ktrans-based perfusion at week 2 (p = 0.018). ConclusionMpMRI parameters differ significantly between hypoxic and non-hypoxic tumor regions, defined on FMISO-PET/CT as gold standard and might represent surrogate markers for tumor hypoxia. These findings suggest that mpMRI may be useful in the future as a surrogate modality for hypoxia imaging in order to personalize CRT.

DCE-MRI of locally-advanced carcinoma of the uterine cervix: Tofts analysis versus non-model-based analyses

BackgroundDynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) may provide biomarkers of the outcome of locally-advanced cervical carcinoma (LACC). There is, however, no agreement on how DCE-MR recordings should be analyzed. Previously, we have analyzed DCE-MRI data of LACC using non-model-based strategies. In the current study, we analyzed DCE-MRI data of LACC using the Tofts pharmacokinetic model, and the biomarkers derived from this analysis were compared with those derived from the non-model-based analyses.MethodsEighty LACC patients given cisplatin-based chemoradiotherapy with curative intent were included in the study. Treatment outcome was recorded as disease-free survival (DFS) and overall survival (OS). DCE-MRI series were analyzed voxelwise to produce Ktrans and ve frequency distributions, and ROC analysis was used to identify the parameters of the frequency distributions having the greatest potential as biomarkers. The prognostic power of these parameters was compared with that of the non-model-based parameters LETV (low-enhancing tumor volume) and TVIS (tumor volume with increasing signal).ResultsPoor DFS and OS were associated with low values of Ktrans, whereas there was no association between treatment outcome and ve. The Ktrans parameters having the greatest prognostic value were p35-Ktrans (the Ktrans value at the 35 percentile of a frequency distribution) and RV-Ktrans (the tumor subvolume with Ktrans values below 0.13 min− 1). Multivariate analysis including clinical parameters and p35-Ktrans or RV-Ktrans revealed that RV-Ktrans was the only independent prognostic factor of DFS and OS. There were significant correlations between RV-Ktrans and LETV and between RV-Ktrans and TVIS, and the prognostic power of RV-Ktrans was similar to that of LETV and TVIS.ConclusionsBiomarkers of the outcome of LACC can be provided by analyzing DCE-MRI series using the Tofts pharmacokinetic model. However, these biomarkers do not appear to have greater prognostic value than biomarkers determined by non-model-based analyses.

Revisiting the identification of tumor sub-volumes predictive of residual uptake after (chemo)radiotherapy: influence of segmentation methods on 18F-FDG PET/CT images

Our aim was to evaluate the impact of the accuracy of image segmentation techniques on establishing an overlap between pre-treatment and post-treatment functional tumour volumes in 18FDG-PET/CT imaging. Simulated images and a clinical cohort were considered. Three different configurations (large, small or non-existent overlap) of a single simulated example was used to elucidate the behaviour of each approach. Fifty-four oesophageal and head and neck (H&N) cancer patients treated with radiochemotherapy with both pre- and post-treatment PET/CT scans were retrospectively analysed. Images were registered and volumes were determined using combinations of thresholds and the fuzzy locally adaptive Bayesian (FLAB) algorithm. Four overlap metrics were calculated. The simulations showed that thresholds lead to biased overlap estimation and that accurate metrics are obtained despite spatially inaccurate volumes. In the clinical dataset, only 17 patients exhibited residual uptake smaller than the pre-treatment volume. Overlaps obtained with FLAB were consistently moderate for esophageal and low for H&N cases across all metrics. Overlaps obtained using threshold combinations varied greatly depending on thresholds and metrics. In both cases overlaps were variable across patients. Our findings do not support optimisation of radiotherapy planning based on pre-treatment 18FDG-PET/CT image definition of high-uptake sub-volumes. Combinations of thresholds may have led to overestimation of overlaps in previous studies.

Spatial Analysis of Tumor Sub-Volume Using Multi-Modality Images for Rhabdomyosarcoma in Children

To compare imaging of multi-modality for evaluating children patients with rhabdomyosacroma. To analyze the spatial relationship of tumor volumes between different imaging techniques. To identify the high risk sub-regions in tumor volumes. 16 pediatric patients with rhabdomyosarcoma underwent imaging with PET/CT, T2-MRI, ADC and spectral CT with iodine contrast. From the spectral CT, the volumetric iodine density and effective Z of patients’ images were extracted. Tumor/GTV contours were generated by clinical physicians on simulation CT. The images of different modalities were rigidly registered on the simulation CTs and the tumor contours were populated to other images in the same coordinate of simulation CTs. The tumor subvolumes of modalities were defined by isothreshold and fuzzy mean clustering. The spatial relations of these sub regions including similarity and incongruity were then evaluated with Dice’s coefficient and ANOVA. The modalities can be categorized with two groups. Group 1: ADC and T2-MRI; Group 2: iodine density, effective Z from spectral CT and PET. In the same patient of most cases, the sub-regions of ADC and T2 had high similarity. The sub-regions of iodine, effective Z and PET had high similarity. The sub-region of ADC showed a disjoint spatial relation with the sub-region of iodine density, e.g. the high iodine density signal of the sub region expressed a low ADC signal in the same region. The fuzzy mean clustering showed the better performance of sub volume determination. The average similarity of between and T2 images has Dice coefficient of 0.88±0.082. The average similarity of between iodine density, effective Z from spectral CT and PET images has Dice coefficient of 0.78±0.042. The average similarity of between ADC and iodine density images has Dice coefficient of 0.14±0.12. The sub-regions of tumor volumes were investigated between images of five modalities. Each modality provides different type of functional imaging based on physical properties of tumor tissues. The spatial correlation of sub-volumes in different modalities can help physicians identify the high risk sub-region of GTV for treatment planning.

Use of Baseline 18F-FDG PET/CT to Identify Initial Sub-Volumes with Local Failure after Concomitant Chemoradiotherapy in Locally Advanced Cervical Cancer

Locally advanced cervical cancer (CC) treated by chemoradiotherapy (CRT) have a significant local recurrence rate. The purpose of this study was to compare the initial high-uptake sub-volume (V1) on baseline18F-FDG PET/CT scans with the metabolic relapse (V2) sites after CRT in locally advanced CC. 18F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) performed before treatment and at relapse in 21 patients diagnosed with LACC and treated with CRT were retrospectively reviewed. CT images at the time of recurrence were registered to baseline CT using the 3D Slicer TM Expert Automated Registration module. The corresponding PET images were then registered using the same shift parameters. The fuzzy locally adaptive Bayesian (FLAB) algorithm was applied using 3 classes (one for background and the other two for tumor) in PET1 to simultaneously define an overall tumor volume and V1. In PET2, FLAB was applied using the standard 2-class version (one for background, one for tumor) in order to define V2. Four indices were used to determine the overlap of V1 and V2 (DICE, overlap fraction, X (X= V1∩V2)/V1) and Y(Y= V1∩V2)/V2)). The mean follow-up ±SD was 26±11 months. The overlap between V1 and V2 were good with an overlap fraction ranging from 0.62-0.81 (0.72±0.05), 0.73-1.00(0.91±0.24), 0.51-1.00 (0.79±0.29), 0.50-1.00 (0.75±0.13), for DICE, overlap fraction, X and Y, respectively. In our study, the overlap between the initial high-uptake sub-volume and the recurrent metabolic tumor sub-volume showed a good concordance. Despite these results need to be investigated in a larger cohort using a more standardized patient repositioning protocol for sequential PET imaging, there are encouraging for potential RT dose escalation in the future.

Mid-treatment MRI Perfusion Changes in Constrictor Muscles Correlate with Dysphagia

Individualized radiation therapy requires assessment of an individual patient’s susceptibility to dysphagia beyond population-based clinical and dosimetric factors. We have previously shown that radiation effects on the pharyngeal constrictors are dose-dependent and that mid-treatment DCE-MRI blood volume (BV) in the constrictors correlate with RT dose. In this study, we aim to evaluate pre- and mid-treatment DCE-MRI imaging parameters in the pharyngeal constrictor muscles and their correlation with dysphagia. Patients with “poor prognosis” head and neck (HN) cancer including human papilloma virus negative (HPV-) loco-regionally advanced HN cancer or stage III AJCC 8 p16+ OPSCC (oropharyngeal squamous cell carcinoma) were enrolled in a randomized phase II trial where high risk tumor subvolumes received 70 vs 86Gy EQD2 with concurrent cisplatin or carboplatin. Multiparametric MRI were acquired at baseline and between fraction (fx) 8-12 of radiation. Analysis were performed on the first 23 patients free of loco-regional recurrence with minimum 12 months of follow-up. Superior, middle and inferior pharyngeal constrictor muscles (SPCM, MPCM and IPCM) were contoured on post contrast T1 MR images and transferred onto BV maps derived from DCE-MRI. Sternocleidomastoid muscle (SCM) was contoured as control. BV changes in the swallowing structures were correlated with RT dose and the following parameters at 3 and 12 months post-RT: IMRT swallow score of videofluoroscopy, EORTC HN35 PRO and any feeding tube use using Spearman rank coefficient. Mean doses to the SPCM, MPCM and IPCM were 59Gy (+/- 6), 56Gy (+/- 9), 33Gy (+/- 17) and respectively. Average BV values were 6.8 (SPCM), 5.8 (MPCM), 5.7 (IPCM) and 2.9 ml/100g (SCM) pretreatment, and changed to 10.4 (SPCM), 9.5 (MPCM), 7.4 (IPCM) and 2.6 (SCM) at 2wk. The increases in BV between pre-treatment and 2wk were significant in SPCM (p=0.01), MPCM (p<0.01), and IPCM (p=0.03), but not in SCM control (p=0.4). There was a significant correlation between the increased BV value at 2wk vs pretreatment and the mean dose in the total constrictor musculature (TPCM) (p=0.03, Spearmen rank correlation). Any feeding tube use was present in 16% of patients at 3 months and 0% at 12 months post-RT. The increased mid-treatment BV values in SPCM and MPCM were significantly correlated with the use of feeding tube at 3 months post-RT (p=0.02 and 0.05, respectively). These findings suggest that MRI BV is a sensitive early indicator of pharyngeal constrictor inflammation and edema. We have also seen significant dispersion of BV estimates between patients, suggesting substantial individual differences in dose-effect relationships. These findings provide a rationale for individualizing swallowing structures doses based on early toxicity prediction. The lack of any feeding tube use at 12mo in this population receiving 70-86Gy to PTVhigh is encouraging.

Early Changes of Diffusion MRI Predict Radiation Therapy Effects in Esophageal Cancer

Concurrent chemoradiation (CRT) is the standard therapy for inoperable locally advanced esophageal cancer (LAEC). However, patients who are insensitive to CRT (20∼30%) may be harmed unnecessarily by the treatment toxicity without survival benefit. An early prediction of treatment response during therapy using non-invasive imaging could identify the patients at high-risk of poor response for individualized adaptive treatment. This study aimed to investigate physiological changes in LAEC using diffusion and perfusion MRI for early predicting treatment response to CRT. Sixty-seven LAEC patients (median age of 64 years) treated with CRT were enrolled in an IRB approved study. All patients received definite thoracic radiotherapy (RT) with a median dose of 60Gy and 30fx, and 73% received concurrent chemotherapy. MRI scans were performed pre-RT and at week 3 during RT (mid-RT), including diffusion weighed images to derive apparent diffusion coefficient (ADC) and dynamic contrast-enhanced images to quantify the fractional blood volume (BV). Gross tumor volume (GTV) was delineated on T2 weighted images. Within the GTV, the hypercellularity volume (VHC) was defined as ADC <1.86x10-3 mm2/s, and the high BV volume (VHBV) as BV >18.2 ml/(100g). ROC curve was applied to assess the prediction abilities of the imaging metrics for CRT outcomes. Associations of the imaging metrics pre-RT and mid-RT and their changes during RT with progression-free survival (PFS) of patients were analyzed by Cox proportional hazards regression analysis. With a median follow-up period of 35.1 months (range 20–41 months), median PFS was 10.1 months. For all patients, the median pre-RT GTV was 28cc±3.2cc, with decreasing of -32.5% ±7.2% to 16.5cc±2.2cc at mid RT. Respectively, VHC and VHBV decreased from 6cc±1.1cc and 4.5cc±1.1cc at pre-RT to 4.1cc±0.7cc and 3.2cc±0.7cc at mid-RT. Smaller VHC at mid-RT (AUC=0.65, p=0.03) and more decreasing in VHC at mid-RT (AUC=0.7, p=0.007) were associated with longer PFS (Table 1). No significant predictive effects were shown in GTV and VHBV. This study suggests that VHC may represent aggressive subvolumes in LAEC. Further analysis will be carried out to correlate the changes in image-phenotype subvolumes with local failure to determine the radiation-resistant tumor subvolumes, which may benefit from treatment intensification.Abstract 1033; Table 1Univariable and multivariable Cox proportional hazards regression analysisCut-off valuesHR95%CIP valueHR95%CIP valueGTVPre-RT33.1cc1.590.88-2.880.13GTVMid-RT13.3cc1.060.63-1.740.89GTVchange-54%1.090.55-2.160.80VHCPre-RT7.3cc1.470.81-2.680.21VHCMid-RT1.1cc2.281.10-4.770.03*2.020.95-4.280.66VHCchange-63%1.950.99-3.820.05*1.680.85-3.330.14VHBVPre-RT10.4cc1.470.94-2.740.14VHBVMid-RT3.3cc1.290.71-2.340.39VHBVchange-54%1.450.75-2.790.27* Statistically significant. Open table in a new tab

Feeding the Data Monster: Data Science in Head and Neck Cancer for Personalized Therapy

ObjectiveHead and neck carcinomas are clinically challenging malignancies because of tumor heterogeneities and resilient tumor subvolumes that require individualized treatment planning and delivery for an improved outcome. Although current approaches to diagnosis and therapy have boosted locoregional control, the long-term survival in this patient group remains unchanged over the last decades. A new approach to head and neck cancer management is therefore needed to better identify patient subgroups that are responsive to specific therapies. The aim of this article is to review the current status of knowledge and practice utilizing big data toward personalized therapy in head and neck cancers based on CT and PET imaging modalities. MethodsLiterature published in English since 2000 was searched using Medline. Additional articles were retrieved via pearling of identified literature. Publications were reviewed and summarized in tabulated format. ResultsStudies based on big data in head and neck cancer are limited; however, the field of radiomics is under continuous development and provides valuable input for personalized treatment. Using PET/PET CT biomarkers for patient treatment individualization and response prediction seems promising, especially in regard to detection of hypoxia and clonogenic cancer stem cells. Literature shows that macroscopic changes in medical images (whether structural or functional) are correlated with biologic and biochemical changes within a tumor. ConclusionCurrent trends in data science suggest that the ideal model for decision support in head and neck cancers should be based on human-machine collaboration, namely, on (1) software-based algorithms, (2) physician innovation collaboratives, and (3) clinician mix optimization.

Clinical Translation of a Quantitative Oxygen Sensor for HDR Cervical Cancer Brachytherapy

Hypoxia is a poor prognostic factor in multiple tumor types. Hypoxic tumors are more resistant to radiation and chemotherapy and are more likely to develop distant metastases. Radiation resistance is of particular interest in cervical cancer, as baseline tumor hypoxia has been shown to be prognostic for local failure and survival in patients treated with primary radiotherapy.1 In vivo studies have shown that radiation resistance may be overcome by increasing the delivered radiation dose by a factor of 2-3. MR-based HDR brachytherapy is well suited for localized dose escalation in hypoxic tumor sub-volumes given the ability to modulate the dose distribution while respecting normal tissue constraints. Currently available clinical techniques for measuring hypoxia are qualitative and limited by poor sensitivity. Meaningful modulation of dose distribution will require an MRI-compatible quantitative oxygen sensor capable of repeated measurements throughout the duration of treatment.

Predicting Gleason Score of Prostate Cancer Patients Using Radiomic Analysis.

Purpose: Use of quantitative imaging features and encoding the intra-tumoral heterogeneity from multi-parametric magnetic resonance imaging (mpMRI) for the prediction of Gleason score is gaining attention as a non-invasive biomarker for prostate cancer (PCa). This study tested the hypothesis that radiomic features, extracted from mpMRI, could predict the Gleason score pattern of patients with PCa.Methods: This analysis included T2-weighted (T2-WI) and apparent diffusion coefficient (ADC, computed from diffusion-weighted imaging) scans of 99 PCa patients from The Cancer Imaging Archive (TCIA). A total of 41 radiomic features were calculated from a local tumor sub-volume (i.e., regions of interest) that is determined by a centroid coordinate of PCa volume, grouped based on their Gleason score patterns. Kruskal-Wallis and Spearman's rank correlation tests were used to identify features related to Gleason score groups. Random forest (RF) classifier model was used to predict Gleason score groups and identify the most important signature among the 41 radiomic features.Results: Gleason score groups could be discriminated based on zone size percentage, large zone size emphasis and zone size non-uniformity values (p < 0.05). These features also showed a significant correlation between radiomic features and Gleason score groups with a correlation value of −0.35, 0.32, 0.42 for the large zone size emphasis, zone size non-uniformity and zone size percentage, respectively (corrected p < 0.05). RF classifier model achieved an average of the area under the curves of the receiver operating characteristic (ROC) of 83.40, 72.71, and 77.35% to predict Gleason score groups (G1) = 6; 6 < (G2) < (3 + 4) and (G3) ≥ 4 + 3, respectively.Conclusion: Our results suggest that the radiomic features can be used as a non-invasive biomarker to predict the Gleason score of the PCa patients.

Automated Vessel Segmentation in Pancreas 4D-MRI using a Novel Transferred Convolutional Neural Network

Pancreatic cancer has dismal prognosis with five-year survival of ∼5%. Although surgery is considered the most effective treatment, only 10-20% of the patients present with resectable disease. Stereotactic body radiation therapy (SBRT) with simultaneous integrated boost (SIB) to tumor sub-volumes abutting major vessels has the potential to sterilize the tumor sub-volumes and down-stage patients for subsequent successful resection. However, abdominal tumor and organs are mobile. SBRT-SIB to small tumor sub-volumes is greatly challenged by ventilation induced tumor and organ motion, making it difficult to define treatment margins. 4D imaging has been used to assess tumor motion in radiation therapy. In a 4D image set, 5 to 10 3D volumetric images are usually generated, making manual segmentation of involved vessels time-consuming and not practical in real clinical setting. The purpose of this study is to apply a novel deep-learning method, two-scale convolutional neural network (CNN) followed by distance and contour regularized level set evolution (DCRLSE), on pancreas 4D-MR data sets to automatically delineate vessels limiting surgery. The 4D-MR images of ten patients were used. Each patient has 10 3D respiration-resolved image sets. For each 3D image set, there are approximately 40 2D axial slices that include vessels limiting surgery. For each patient’s 4D-MRI data-set, 300 image frames were analyzed. The proposed transferred CNN routine is composed of two tasks. In the source task, a relatively large amount of data is used to train the paired CNN. In the target task, a relatively small amount of data is used to fine-tune the network. In the test phase, Paired CNN is used to locate and roughly segment the vessel. Small holes in the segmentation are filled using morphological operation. The DCRLSE model is then used to refine the contours. Transferred learning is used to improve the performance by exploiting the similarity among different patients. After the binary mask is trained on one patient, the network weights are transferred to initialize the CNN training of the second patient’s network. The weights are fine-tuned on selected breathing phases to find the minimal number of images needed for the new segmentation task. A widely used segmentation method, level-set, was used for comparison. Manual segmentation was used as ground truth. The segmentation results were evaluated visually, as well as quantitatively using four metrics: sensitivity, specificity, dice similarity coefficient (DSC) and modified Hausdorff distance (HD). CNN-DCRLSE method visually showed more consistent segmentation results comparing to level-set method. It was able to achieve average sensitivity of 0.876±0.070 (mean±σ), specificity of 0.999±0.001, DSC of 0.863±0.081, and HD (mm) of 0.099±0.059. The CNN-DCRLSE method was successfully implemented on 4D-MR images to automatically segment major blood vessels that limit surgery for pancreatic cancer patients.