Automatic gross tumor segmentation of canine head and neck cancer using deep learning and cross-species transfer learning.

Background: Radiotherapy (RT) is increasingly being used on dogs with spontaneous head and neck cancer (HNC), which account for a large percentage of veterinary patients treated with RT. Accurate definition of the gross tumor volume (GTV) is a vital part of RT planning, ensuring adequate dose coverage of the tumor while limiting the radiation dose to surrounding tissues. Currently the GTV is contoured manually in medical images, which is a time-consuming and challenging task. Purpose: The purpose of this study was to evaluate the applicability of deep learning-based automatic segmentation of the GTV in canine patients with HNC. Materials and methods: Contrast-enhanced computed tomography (CT) images and corresponding manual GTV contours of 36 canine HNC patients and 197 human HNC patients were included. A 3D U-Net convolutional neural network (CNN) was trained to automatically segment the GTV in canine patients using two main approaches: (i) training models from scratch based solely on canine CT images, and (ii) using cross-species transfer learning where models were pretrained on CT images of human patients and then fine-tuned on CT images of canine patients. For the canine patients, automatic segmentations were assessed using the Dice similarity coefficient (Dice), the positive predictive value, the true positive rate, and surface distance metrics, calculated from a four-fold cross-validation strategy where each fold was used as a validation set and test set once in independent model runs. Results: CNN models trained from scratch on canine data or by using transfer learning obtained mean test set Dice scores of 0.55 and 0.52, respectively, indicating acceptable auto-segmentations, similar to the mean Dice performances reported for CT-based automatic segmentation in human HNC studies. Automatic segmentation of nasal cavity tumors appeared particularly promising, resulting in mean test set Dice scores of 0.69 for both approaches. Conclusion: In conclusion, deep learning-based automatic segmentation of the GTV using CNN models based on canine data only or a cross-species transfer learning approach shows promise for future application in RT of canine HNC patients.

Lack of radioprotection knowledge and compliance in Norwegian equine ambulatory practice.

There is an increasing use of mobile radiographic units in equine ambulatory practices in Norway. Horse owners or handlers often participate in the radiographic examination in a non-controlled area. The aim of this descriptive, cross-sectional, survey study was to evaluate the radiation safety and protection in use of mobile radiography, and to identify areas where special attention from the regulatory authorities as well as veterinary educators would be required. A questionnaire was distributed to all equine veterinarians assumed to have access to mobile radiographic units, as part of a formal inspection in cooperation with the Norwegian Radiation Protection Authority. Few practices met the regulatory requirements of notifying the authorities of their radiographic units and designation of a radiation protection officer. The minority of the practices performed periodic quality assurance of their equipment. Many of the practices performed all of their radiographic examinations off-site. The examinations were most often performed in the aisle outside the horses' stalls, and few practices established an operating zone. The horse owner or handler participated in the radiographic examination in almost all of the practices. Few practices used dosimeters for determination of the radiation exposure. The study shows that there are major deficits in regulatory compliance in ambulatory equine radiography practices in Norway. The study also suggests that less stringent regulations and supervision may translate into less stringent radioprotection practices.

Dose or 'LET' painting--What is optimal in particle therapy of hypoxic tumors?

BACKGROUND: Dose painting is a concept that may increase the tumor control probability (TCP). In particle therapy of hypoxic tumors, it may also be beneficial to redistribute the linear energy transfer (LET) so that the oxygen effect is minimized; so-called LET painting. The purpose of the present study was to use TCP estimates for comparing dose and LET painting of hypoxic tumors. MATERIAL AND METHODS: Protons, lithium ions and carbon ions were considered. Tumor images tentatively depicting hypoxia were used as input. Optimal dose prescription maps were obtained by optimizing TCP under dose and/or LET redistribution. TCPs were compared to those resulting from conventional particle therapy with no dose or LET painting. The therapeutic gain at a given iso-effect was calculated. Treatment adaptation during therapy in response to changes in the spatial hypoxia distribution was also considered. RESULTS: Both dose and LET painting gave higher TCPs compared to conventional particle therapy, irrespective of particle type. The therapeutic gain from LET painting, dose painting and combined dose+ LET painting was 1.09/1.43/1.45, 1.24/1.32/1.37 and 1.16/1.23/1.28 for protons, lithium ions and carbon ions, respectively. The importance of treatment adaptation was less pronounced for particles heavier than protons. CONCLUSION: Dose painting results in higher TCP than LET painting, in particular for protons. For heavier ions, LET painting may also give an enhanced tumor effect compared to conventional particle therapy. Combined dose+ LET painting may only give a marginally increased effect compared to dose painting only. Adaptive carbon ion dose painting seems to be of less importance.

Functional imaging to monitor vascular and metabolic response in canine head and neck tumors during fractionated radiotherapy.

Radiotherapy causes alterations in tumor biology, and non-invasive early assessment of such alterations may become useful for identifying treatment resistant disease. The purpose of the current work is to assess changes in vascular and metabolic features derived from functional imaging of canine head and neck tumors during fractionated radiotherapy. Material and methods. Three dogs with spontaneous head and neck tumors received intensity-modulated radiotherapy (IMRT). Contrast-enhanced cone beam computed tomography (CE-CBCT) at the treatment unit was performed at five treatment fractions. Dynamic (18)FDG-PET (D-PET) was performed prior to the start of radiotherapy, at mid-treatment and at 3-12 weeks after the completion of treatment. Tumor contrast enhancement in the CE-CBCT images was used as a surrogate for tumor vasculature. Vascular and metabolic tumor parameters were further obtained from the D-PET images. Changes in these tumor parameters were assessed, with emphasis on intra-tumoral distributions. Results. For all three patients, metabolic imaging parameters obtained from D-PET decreased from the pre- to the inter-therapy session. Correspondingly, for two of three patients, vascular imaging parameters obtained from both CE-CBCT and D-PET increased. Only one of the tumors showed a clear metabolic response after therapy. No systematic changes in the intra-tumor heterogeneity in the imaging parameters were found. Conclusion. Changes in vascular and metabolic parameters could be detected by the current functional imaging methods. Vascular tumor features from CE-CBCT and D-PET corresponded well. CE-CBCT is a potential method for easy response assessment when the patient is at the treatment unit.

Spatiotemporal analysis of tumor uptake patterns in dynamic (18)FDG-PET and dynamic contrast enhanced CT.

BACKGROUND: Molecular and functional imaging techniques such as dynamic positron emission tomography (DPET) and dynamic contrast enhanced computed tomography (DCECT) may provide improved characterization of tumors compared to conventional anatomic imaging. The purpose of the current work was to compare spatiotemporal uptake patterns in DPET and DCECT images. MATERIALS AND METHODS: A PET/CT protocol comprising DCECT with an iodine based contrast agent and DPET with (18)F-fluorodeoxyglucose was set up. The imaging protocol was used for examination of three dogs with spontaneous tumors of the head and neck at sessions prior to and after fractionated radiotherapy. Software tools were developed for downsampling the DCECT image series to the PET image dimensions, for segmentation of tracer uptake pattern in the tumors and for spatiotemporal correlation analysis of DCECT and DPET images. RESULTS: DCECT images evaluated one minute post injection qualitatively resembled the DPET images at most imaging sessions. Segmentation by region growing gave similar tumor extensions in DCECT and DPET images, with a median Dice similarity coefficient of 0.81. A relatively high correlation (median 0.85) was found between temporal tumor uptake patterns from DPET and DCECT. The heterogeneity in tumor uptake was not significantly different in the DPET and DCECT images. The median of the spatial correlation was 0.72. CONCLUSIONS: DCECT and DPET gave similar temporal wash-in characteristics, and the images also showed a relatively high spatial correlation. Hence, if the limited spatial resolution of DPET is considered adequate, a single DPET scan only for assessing both tumor perfusion and metabolic activity may be considered. However, further work on a larger number of cases is needed to verify the correlations observed in the present study.

Dosimetric verification of biologically adapted IMRT.

PURPOSE: To investigate the delivery of biologically adapted high resolution intensity modulated radiotherapy (IMRT) to an anthropomorphic phantom, using dosimetric and radiobiologic measures. METHODS: A compartment based 3D hypoxia map of a highly heterogeneous tumor was imported into the CT planning basis of an anthropomorphic phantom. Biologically adapted IMRT was planned according to the corresponding 3D dose prescription map with elevated dose to the hypoxic regions. Three treatment fractions were delivered to the phantom by means of a conventional linear accelerator equipped with a high resolution micro multileaf collimator (mMLC). EDR2 radiographic films were positioned in two planes of the phantom during irradiation. Software was developed to analyze dose distributions from the prescription, dose plan, and films. Dose distributions were scored within each of four radiobiologic tumor compartments. The treatment effect in each tumor compartment was estimated as the equivalent uniform dose (EUD). A conventional gamma analysis was utilized for quantitative comparisons of planned and delivered dose distributions. RESULTS: The planned and delivered dose maps qualitatively resembled the prescribed dose map. The gamma analyzes showed that, on average for all films, more than 95% of the pixels within the tumor passed the 3%/2 mm criteria. For compartments with increasing degree of hypoxia and thereby increased prescribed dose, the planned and delivered EUDs were severely reduced compared to the prescription. The prescribed compartmental doses were met only for the most oxic compartment. The mean tumor dose as measured by the films was 6.6% lower than the corresponding planned dose. CONCLUSIONS: Biologically adapted radiotherapy may be delivered with high precision according to the dose plan. However, the large reduction in compartmental EUD values from prescribed to planned treatment indicated lower tumor effect than expected for such a treatment.

Adaptive radiotherapy based on contrast enhanced cone beam CT imaging.

Cone beam CT (CBCT) imaging has become an integral part of radiation therapy, with images typically used for offline or online patient setup corrections based on bony anatomy co-registration. Ideally, the co-registration should be based on tumor localization. However, soft tissue contrast in CBCT images may be limited. In the present work, contrast enhanced CBCT (CECBCT) images were used for tumor visualization and treatment adaptation. Material and methods. A spontaneous canine maxillary tumor was subjected to repeated cone beam CT imaging during fractionated radiotherapy (10 fractions in total). At five of the treatment fractions, CECBCT images, employing an iodinated contrast agent, were acquired, as well as pre-contrast CBCT images. The tumor was clearly visible in post-contrast minus pre-contrast subtraction images, and these contrast images were used to delineate gross tumor volumes. IMRT dose plans were subsequently generated. Four different strategies were explored: 1) fully adapted planning based on each CECBCT image series, 2) planning based on images acquired at the first treatment fraction and patient repositioning following bony anatomy co-registration, 3) as for 2), but with patient repositioning based on co-registering contrast images, and 4) a strategy with no patient repositioning or treatment adaptation. The equivalent uniform dose (EUD) and tumor control probability (TCP) calculations to estimate treatment outcome for each strategy. Results. Similar translation vectors were found when bony anatomy and contrast enhancement co-registration were compared. Strategy 1 gave EUDs closest to the prescription dose and the highest TCP. Strategies 2 and 3 gave EUDs and TCPs close to that of strategy 1, with strategy 3 being slightly better than strategy 2. Even greater benefits from strategies 1 and 3 are expected with increasing tumor movement or deformation during treatment. The non-adaptive strategy 4 was clearly inferior to all three adaptive strategies. Conclusion. CECBCT may prove useful for adaptive radiotherapy.

Influence of MLC leaf width on biologically adapted IMRT plans.

INTRODUCTION: High resolution beam delivery may be required for optimal biology-guided adaptive therapy. In this work, we have studied the influence of multi leaf collimator (MLC) leaf widths on the treatment outcome following adapted IMRT of a hypoxic tumour. MATERIAL AND METHODS: Dynamic contrast enhanced MR images of a dog with a spontaneous tumour in the nasal region were used to create a tentative hypoxia map following a previously published procedure. The hypoxia map was used as a basis for generating compartmental gross tumour volumes, which were utilised as planning structures in biologically adapted IMRT. Three different MLCs were employed in inverse treatment planning, with leaf widths of 2.5 mm, 5 mm and 10 mm. The number of treatment beams and the degree of step-and-shoot beam modulation were varied. By optimising the tumour control probability (TCP) function, optimal compartmental doses were derived and used as target doses in the inverse planning. Resulting IMRT dose distributions and dose volume histograms (DVHs) were exported and analysed, giving estimates of TCP and compartmental equivalent uniform doses (EUDs). The impact of patient setup accuracy was simulated. RESULTS: The MLC with the smallest leaf width (2.5 mm) consistently gave the highest TCPs and compartmental EUDs, assuming no setup error. The difference between this MLC and the 5 mm MLC was rather small, while the MLC with 10 mm leaf width gave considerably lower TCPs. When including random and systematic setup errors, errors larger than 5 mm gave only small differences between the MLC types. For setup errors larger than 7 mm no differences were found between non-uniform and uniform dose distributions. CONCLUSIONS: Biologically adapted radiotherapy may require MLCs with leaf widths smaller than 10 mm. However, for a high probability of cure it is crucial that accurate patient setup is ensured.

Adapting biological feedback in radiotherapy.

Recently, there has been much interest in how to use information on patient-specific tumor biology and normal tissue function to individualize cancer treatment. In radiation therapy, dose may be escalated to radioresistant regions within a tumor, or regions of particular functional importance in normal organs may be preferentially spared. However, tumor and normal tissue biology may change during treatment, and adaptation of therapy may be necessary to ensure that optimal therapy is delivered. Furthermore, changes in tumor and normal tissue biology during early treatment may be predictive for the outcome of radiotherapy, and this information could be used for individual adaptation of the remaining part of the treatment. In the present study, we address variations that may occur in tumor and normal tissue radiobiological properties during radiotherapy, and how these may be related to the response to treatment. Moreover, we discuss the criteria for when to adapt treatment and how this adaptation should be performed. Finally, we discuss to what degree biologically adapted radiotherapy may be expected to improve treatment outcome and which issues need to be resolved for this strategy to reach its full potential.

Strategies for biologic image-guided dose escalation: a review.

There is increasing interest in how to incorporate functional and molecular information obtained by noninvasive, three-dimensional tumor imaging into radiotherapy. The key issues are to identify radioresistant regions that can be targeted for dose escalation, and to develop radiation dose prescription and delivery strategies providing optimal treatment for the individual patient. In the present work, we review the proposed strategies for biologic image-guided dose escalation with intensity-modulated radiation therapy. Biologic imaging modalities and the derived images are discussed, as are methods for target volume delineation. Different dose escalation strategies and techniques for treatment delivery and treatment plan evaluation are also addressed. Furthermore, we consider the need for response monitoring during treatment. We conclude with a summary of the current status of biologic image-based dose escalation and of areas where further work is needed for this strategy to become incorporated into clinical practice.

DCEMRI monitoring of canine tumors during fractionated radiotherapy.

PURPOSE: To monitor the contrast enhancement in spontaneous canine tumors during fractionated radiotherapy by Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCEMRI). METHODS AND MATERIAL: Six dogs with tumors in the oral or nasal cavity received fractionated conformal radiotherapy with 54 Gy given in 18 fractions. T1 weighted DCE imaging was performed prior to each treatment fraction. The tumor was manually delineated in the MR images following every imaging session, and the time dependence of the Relative Signal Intensity (RSI) in the tumor was extracted voxel by voxel. RSI images at the time of maximum enhancement were generated, in addition to images of the initial slope of the RSI curves. The dependence of the median RSI and median slope in the tumor on the accumulated radiation dose was investigated, and images obtained at different treatment fraction were compared by correlation analysis. RESULTS: Five of the six tumors regressed during treatment. The dose dependence of the RSI varied between the tumors, with some showing an increase and others a decrease in RSI with dose. This was also the case for the initial slope of the RSI curves. The correlation between images acquired before the first treatment fraction and subsequent fractions was in general low, and decreased significantly with accumulated radiation dose for five of six tumors. CONCLUSIONS: Large individual variations in the dose response of tumor contrast enhancement were found. Decreasing image correlation resulted both from tumor regression and intratumoral changes in the RSI distribution during treatment. These findings may have consequences for treatment design in biological image-guided radiotherapy.

On the parameter describing the generalised equivalent uniform dose (gEUD) for tumours.

The purpose of the present work was to estimate the parameter 'a' describing the generalised equivalent uniform dose (gEUD) for tumours and its dependence on radiobiological parameters. The consequences of uncertainties in a on the gEUD were also studied. An estimate of a was found by requiring that, for a given target dose distribution, the mechanistic EUD (based on radiobiological linear quadratic modelling) equals gEUD. The estimate of a was found to depend on the dose distribution, and decreased with factors that increase the slope of the cell survival curve (i.e. decreasing alpha/beta values and increasing alpha values). Furthermore, the parameter a was estimated for 35 prostate cancer IMRT plans of varying dose distributions, for two sets of previously published radiobiological parameters: (1) alpha=0.15 Gy(-1) and alpha/beta=3 Gy, and (2) alpha=0.26 Gy(-1) and alpha/beta=10 Gy. The estimated values of a ranged from -25.6 to -22.4 for all combinations of dose distributions and parameter sets. Uncertainties in a were found to give only small uncertainties in gEUD. Although the current work shows limitations of the gEUD model for tumours, gEUD may still be preferable for biological treatment plan optimization, evaluation and reporting.

Radiotherapy adapted to spatial and temporal variability in tumor hypoxia.

PURPOSE: To explore the feasibility and clinical potential of adapting radiotherapy to temporal and spatial variations in tumor oxygenation. METHODS AND MATERIALS: Repeated dynamic contrast enhanced magnetic resonance (DCEMR) images were taken of a canine sarcoma during the course of fractionated radiation therapy. The tumor contrast enhancement was assumed to represent the oxygen distribution. The IMRT plans were retrospectively adapted to the DCEMR images by employing tumor dose redistribution. Optimized nonuniform tumor dose distributions were calculated and compared with a uniform dose distribution delivering the same integral dose to the tumor. Clinical outcome was estimated from tumor control probability (TCP) and normal tissue complication probability (NTCP) modeling. RESULTS: The biologically adapted treatment was found to give a substantial increase in TCP compared with conventional radiotherapy, even when only pretreatment images were used as basis for the treatment planning. The TCP was further increased by repeated replanning during the course of treatment, and replanning twice a week was found to give near optimal TCP. Random errors in patient positioning were found to give a small decrease in TCP, whereas systematic errors were found to reduce TCP substantially. NTCP for the adapted treatment was similar to or lower than for the conventional treatment, both for parallel and serial normal tissue structures. CONCLUSION: Biologically adapted radiotherapy is estimated to improve treatment outcome of tumors having spatial and temporal variations in radiosensitivity.

Optimization of tumour control probability in hypoxic tumours by radiation dose redistribution: a modelling study.

Tumour hypoxia is a known cause of clinical resistance to radiation therapy. The purpose of this work was to model the effects on tumour control probability (TCP) of selectively boosting the dose to hypoxic regions in a tumour, while keeping the mean tumour dose constant. A tumour model with a continuous oxygen distribution, incorporating pO(2) histograms published for head and neck patients, was developed. Temporal and spatial variations in the oxygen distribution, non-uniform cell density and cell proliferation during treatment were included in the tumour modelling. Non-uniform dose prescriptions were made based on a segmentation of the tumours into four compartments. The main findings were: (1) Dose redistribution considerably improved TCP for all tumours. (2) The effect on TCP depended on the degree of reoxygenation during treatment, with a maximum relative increase in TCP for tumours with poor or no reoxygenation. (3) Acute hypoxia reduced TCP moderately, while underdosing chronic hypoxic cells gave large reductions in TCP. (4) Restricted dose redistribution still gave a substantial increase in TCP as compared to uniform dose boosts. In conclusion, redistributing dose according to tumour oxygenation status might increase TCP when the tumour response to radiotherapy is limited by chronic hypoxia. This could potentially improve treatment outcome in a subpopulation of patients who respond poorly to conventional radiotherapy.

Adapting radiotherapy to hypoxic tumours.

In the current work, the concepts of biologically adapted radiotherapy of hypoxic tumours in a framework encompassing functional tumour imaging, tumour control predictions, inverse treatment planning and intensity modulated radiotherapy (IMRT) were presented. Dynamic contrast enhanced magnetic resonance imaging (DCEMRI) of a spontaneous sarcoma in the nasal region of a dog was employed. The tracer concentration in the tumour was assumed related to the oxygen tension and compared to Eppendorf histograph measurements. Based on the pO(2)-related images derived from the MR analysis, the tumour was divided into four compartments by a segmentation procedure. DICOM structure sets for IMRT planning could be derived thereof. In order to display the possible advantages of non-uniform tumour doses, dose redistribution among the four tumour compartments was introduced. The dose redistribution was constrained by keeping the average dose to the tumour equal to a conventional target dose. The compartmental doses yielding optimum tumour control probability (TCP) were used as input in an inverse planning system, where the planning basis was the pO(2)-related tumour images from the MR analysis. Uniform (conventional) and non-uniform IMRT plans were scored both physically and biologically. The consequences of random and systematic errors in the compartmental images were evaluated. The normalized frequency distributions of the tracer concentration and the pO(2) Eppendorf measurements were not significantly different. 28% of the tumour had, according to the MR analysis, pO(2) values of less than 5 mm Hg. The optimum TCP following a non-uniform dose prescription was about four times higher than that following a uniform dose prescription. The non-uniform IMRT dose distribution resulting from the inverse planning gave a three times higher TCP than that of the uniform distribution. The TCP and the dose-based plan quality depended on IMRT parameters defined in the inverse planning procedure (fields and step-and-shoot intensity levels). Simulated random and systematic errors in the pO(2)-related images reduced the TCP for the non-uniform dose prescription. In conclusion, improved tumour control of hypoxic tumours by dose redistribution may be expected following hypoxia imaging, tumour control predictions, inverse treatment planning and IMRT.