DeepDose: a robust deep learning-based dose engine for abdominal tumours in a 1.5 T MRI radiotherapy system.

We present a robust deep learning-based framework for dose calculations of abdominal tumours in a 1.5 T MRI radiotherapy system. For a set of patient plans, a convolutional neural network is trained on the dose of individual multi-leaf-collimator segments following the DeepDose framework. It can then be used to predict the dose distribution per segment for a set of patient anatomies. The network was trained using data from three anatomical sites of the abdomen: prostate, rectal and oligometastatic tumours. A total of 216 patient fractions were used, previously treated in our clinic with fixed-beam IMRT using the Elekta MR-linac. For the purpose of training, 176 fractions were used with random gantry angles assigned to each segment, while 20 fractions were used for the validation of the network. The ground truth data were calculated with a Monte Carlo dose engine at 1% statistical uncertainty per segment. For a total of 20 independent abdominal test fractions with the clinical angles, the network was able to accurately predict the dose distributions, achieving 99.4% ± 0.6% for the whole plan prediction at the 3%/3 mm gamma test. The average dose difference and standard deviation per segment was 0.3% ± 0.7%. Additional dose prediction on one cervical and one pancreatic case yielded high dose agreement of 99.9% and 99.8% respectively for the 3%/3 mm criterion. Overall, we show that our deep learning-based dose engine calculates highly accurate dose distributions for a variety of abdominal tumour sites treated on the MR-linac, in terms of performance and generality.

Assessment of Secondary Sarcomas Among Patients With Cancer of the Abdomen or Pelvis Who Received Combinations of Surgery, Radiation, and Chemotherapy vs Surgery Alone.

Importance: The association of radiation and chemotherapy with the development of secondary sarcoma is known, but the contemporary risk has not been well characterized for patients with cancers of the abdomen and pelvis. Objective: To compare the risk of secondary sarcoma among patients treated with combinations of surgery, radiation, or chemotherapy with patients treated with surgery alone and the general population. Design, Setting, and Participants: This population-based cohort study included 173 580 patients in Ontario, Canada, with nonmetastatic cancer of the prostate, bladder, colon, rectum or anus, cervix, uterus, or testis. Patients were enrolled from January 1, 2002, to January 31, 2017. Data analysis was conducted from March 1, 2019, to January 31, 2020. Exposures: Treatment combinations of radiation, chemotherapy, and surgery. Main Outcome and Measures: Diagnosis of sarcoma based on histologic codes from the Ontario Cancer Registry. Time to sarcoma was compared using a cause-specific proportional hazard model. Results: Of 173 580 patients, most were men (125 080 [72.1%]), and the largest group was aged between 60 and 69 years (58 346 [33.6%]). Most patients had genitourinary cancer (86 235 [51.4%]) or colorectal cancer (69 241 [39.9%]). Overall, 64 301 (37.1%) received surgery alone, 51 220 (29.5%) received radiation alone, 15 624 (9.0%) were treated with radiation and chemotherapy, 15 252 (8.8%) received radiation with surgery, and 11 822 (6.8%) received all 3 treatments. A total of 332 patients (0.2%) had sarcomas develop during a median (interquartile range) follow-up of 5.7 (2.2-8.9) years. The incidence of sarcoma was 0.3% among those who underwent radiation alone (138 of 51 220) and radiation with chemotherapy (40 of 15 624), 0.2% among those who received radiation and surgery (36 of 15 252) and all 3 modalities (25 of 11 822), and 0.1% among those who received surgery with chemotherapy (13 of 14 861) and surgery alone (80 of 64 801). Compared with a reference group of patients who had surgery alone, the greatest risk of sarcoma was found among patients who underwent a combination of radiation and chemotherapy (cause-specific relative hazard [csRH], 4.07; 95% CI, 2.75-6.01; P < .001), followed by patients who had radiation alone (csRH, 2.35; 95% CI, 1.77-3.12; P < .001), radiation with surgery (csRH, 2.33; 95% CI, 1.57-3.46; P < .001), and all 3 modalities (csRH, 2.27; 95% CI, 1.44-3.58; P < .001). In the general population, 7987 events occurred during 46 554 803 person-years (17.2 events per 100 000 person-years). The standardized incidence ratio for sarcoma among patients treated with radiation compared with the general population was 2.41 (95% CI, 1.57-3.69; 41.3 events per 100 000 person-years). The annual number of cases of sarcoma increased from 2009 (15 per 100 000 persons) to 2016 (32 per 100 000 persons), but the annual rate did not change during the study period. Conclusions and Relevance: In this cohort study, patients treated with radiation or chemotherapy for abdominopelvic cancers had an increased rate of sarcoma. Although the absolute rate is low, patients and physicians should be aware of this increased risk of developing sarcoma.

Proton beams in cancer treatments: Clinical outcomes and dosimetric comparisons with photon therapy.

PURPOSE: To review current evidence of the role of proton therapy (PT) in other tumors than skull base, sinusal/parasinusal, spinal and pediatric tumors; to determine medico-economic aspects raised by PT. MATERIAL AND METHODS: A systematic review on Medline was performed with the following keywords: proton therapy, proton beam, protontherapy, cancer; publications with comparison between PT and photon-therapy were also selected. RESULTS: In silico studies have shown superiority (better dose delivery to the target and/or to organs at risk) of PT toward photon-therapy in most of thoracic and abdominal malignant tumors. Potential benefits of PT could be: reduction of toxicities (including radiation-induced cancer), increase of tumor control through a dose-escalation approach, hypofractionation. Cost of treatment is always cited as an issue which actually can be managed by a precise patient selection making PT a cost-effective procedure. Comparison plan with photon therapy may be useful to determine the dosimetric and clinical advantages of PT (Normal Tissue Complications Probability). CONCLUSION: PT may be associated with a great advantage compared to the best photon-therapies in various types of cancers. Accumulation of clinical data is on-going and will challenge the in silico data analysis. Some indications are associated with strong superiority of PT and may be discussed as a new standard within prospective observational studies.

Quality of Intensity Modulated Radiation Therapy Treatment Plans Using a 6°Co Magnetic Resonance Image Guidance Radiation Therapy System.

PURPOSE: This work describes a commercial treatment planning system, its technical features, and its capabilities for creating (60)Co intensity modulated radiation therapy (IMRT) treatment plans for a magnetic resonance image guidance radiation therapy (MR-IGRT) system. METHODS AND MATERIALS: The ViewRay treatment planning system (Oakwood Village, OH) was used to create (60)Co IMRT treatment plans for 33 cancer patients with disease in the abdominal, pelvic, thorax, and head and neck regions using physician-specified patient-specific target coverage and organ at risk (OAR) objectives. Backup plans using a third-party linear accelerator (linac)-based planning system were also created. Plans were evaluated by attending physicians and approved for treatment. The (60)Co and linac plans were compared by evaluating conformity numbers (CN) with 100% and 95% of prescription reference doses and heterogeneity indices (HI) for planning target volumes (PTVs) and maximum, mean, and dose-volume histogram (DVH) values for OARs. RESULTS: All (60)Co IMRT plans achieved PTV coverage and OAR sparing that were similar to linac plans. PTV conformity for (60)Co was within <1% and 3% of linac plans for 100% and 95% prescription reference isodoses, respectively, and heterogeneity was on average 4% greater. Comparisons of OAR mean dose showed generally better sparing with linac plans in the low-dose range <20 Gy, but comparable sparing for organs with mean doses >20 Gy. The mean doses for all (60)Co plan OARs were within clinical tolerances. CONCLUSIONS: A commercial (60)Co MR-IGRT device can produce highly conformal IMRT treatment plans similar in quality to linac IMRT for a variety of disease sites. Additional work is in progress to evaluate the clinical benefit of other novel features of this MR-IGRT system.

Patient-specific quality assurance for the delivery of (60)Co intensity modulated radiation therapy subject to a 0.35-T lateral magnetic field.

PURPOSE: This work describes a patient-specific dosimetry quality assurance (QA) program for intensity modulated radiation therapy (IMRT) using ViewRay, the first commercial magnetic resonance imaging-guided RT device. METHODS AND MATERIALS: The program consisted of: (1) a 1-dimensional multipoint ionization chamber measurement using a customized 15-cm(3) cube-shaped phantom; (2) 2-dimensional (2D) radiographic film measurement using a 30- × 30- × 20-cm(3) phantom with multiple inserted ionization chambers; (3) quasi-3D diode array (ArcCHECK) measurement with a centrally inserted ionization chamber; (4) 2D fluence verification using machine delivery log files; and (5) 3D Monte Carlo (MC) dose reconstruction with machine delivery files and phantom CT. RESULTS: Ionization chamber measurements agreed well with treatment planning system (TPS)-computed doses in all phantom geometries where the mean ± SD difference was 0.0% ± 1.3% (n=102; range, -3.0%-2.9%). Film measurements also showed excellent agreement with the TPS-computed 2D dose distributions where the mean passing rate using 3% relative/3 mm gamma criteria was 94.6% ± 3.4% (n=30; range, 87.4%-100%). For ArcCHECK measurements, the mean ± SD passing rate using 3% relative/3 mm gamma criteria was 98.9% ± 1.1% (n=34; range, 95.8%-100%). 2D fluence maps with a resolution of 1 × 1 mm(2) showed 100% passing rates for all plan deliveries (n=34). The MC reconstructed doses to the phantom agreed well with planned 3D doses where the mean passing rate using 3% absolute/3 mm gamma criteria was 99.0% ± 1.0% (n=18; range, 97.0%-100%), demonstrating the feasibility of evaluating the QA results in the patient geometry. CONCLUSIONS: We developed a dosimetry program for ViewRay's patient-specific IMRT QA. The methodology will be useful for other ViewRay users. The QA results presented here can assist the RT community to establish appropriate tolerance and action limits for ViewRay's IMRT QA.

Radiotherapy of abdomen with precise renal assessment with SPECT/CT imaging (RAPRASI): design and methodology of a prospective trial to improve the understanding of kidney radiation dose response.

BACKGROUND: The kidneys are a principal dose-limiting organ in radiotherapy for upper abdominal cancers. The current understanding of kidney radiation dose response is rudimentary. More precise dose-volume response models that allow direct correlation of delivered radiation dose with spatio-temporal changes in kidney function may improve radiotherapy treatment planning for upper-abdominal tumours. METHODS/DESIGN: The Radiotherapy of Abdomen with Precise Renal Assessment with SPECT/CT Imaging (RAPRASI) is an observational clinical research study with participating sites at Sir Charles Gairdner Hospital (SCGH) in Perth, Australia and the Peter MacCallum Cancer Centre (PMCC) in Melbourne, Australia. Eligible patients are those with upper gastrointestinal cancer, without metastatic disease, undergoing conformal radiotherapy that will involve incidental radiation to one or both kidneys. For each patient, total kidney function is being assessed before commencement of radiotherapy treatment and then at 4, 12, 26, 52 and 78 weeks after the first radiotherapy fraction, using two procedures: a Glomerular Filtration Rate (GFR) measurement using the 51Cr-ethylenediamine tetra-acetic acid (EDTA) clearance; and a regional kidney perfusion measurement assessing renal uptake of 99mTc-dimercaptosuccinic acid (DMSA), imaged with a Single Photon Emission Computed Tomography / Computed Tomography (SPECT/CT) system. The CT component of the SPECT/CT provides the anatomical reference of the kidney's position. The data is intended to reveal changes in regional kidney function over the study period after the radiotherapy. These SPECT/CT scans, co-registered with the radiotherapy treatment plan, will provide spatial correlation between the radiation dose and regional renal function as assessed by SPECT/CT. From this correlation, renal response patterns will likely be identified with the purpose of developing a predictive model. TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry: ACTRN12609000322235.

Novel assessment of renal motion in children as measured via four-dimensional computed tomography.

OBJECTIVES: Abdominal intensity-modulated radiation therapy and proton therapy require quantification of target and organ motion to optimize localization and treatment. Although addressed in adults, there is no available literature on this issue in pediatric patients. We assessed physiologic renal motion in pediatric patients. METHODS AND MATERIALS: Twenty free-breathing pediatric patients at a median age of 8 years (range, 2-18 years) with intra-abdominal tumors underwent computed tomography simulation and four-dimensional computed tomography acquisition (slice thickness, 3 mm). Kidneys and diaphragms were contoured during eight phases of respiration to estimate center-of-mass motion. We quantified center of kidney mass mobility vectors in three dimensions: anteroposterior (AP), mediolateral (ML), and superoinferior (SI). RESULTS: Kidney motion decreases linearly with decreasing age and height. The 95% confidence interval for the averaged minima and maxima of renal motion in children younger than 9 years was 5-9 mm in the ML direction, 4-11 mm in the AP direction, and 12-25 mm in the SI dimension for both kidneys. In children older than 9 years, the same confidence interval reveals a widening range of motion that was 5-16 mm in the ML direction, 6-17 mm in the AP direction, and 21-52 mm in the SI direction. Although not statistically significant, renal motion correlated with diaphragm motion in older patients. The correlation between diaphragm motion and body mass index was borderline (r = 0.52, p = 0.0816) in younger patients. CONCLUSIONS: Renal motion is age and height dependent. Measuring diaphragmatic motion alone does not reliably quantify pediatric renal motion. Renal motion in young children ranges from 5 to 25 mm in orientation-specific directions. The vectors of motion range from 5 to 52 mm in older children. These preliminary data represent novel analyses of pediatric intra-abdominal organ motion.

Accuracy of proton beam range verification using post-treatment positron emission tomography/computed tomography as function of treatment site.

PURPOSE: For 23 patients, an off-line positron emission tomography scan and a computed tomography scan after proton radiotherapy was performed at the Massachusetts General Hospital to assess in vivo treatment verification. A well-balanced population of patients was investigated to assess the effect of the tumor location on the accuracy of the technique. METHODS AND MATERIALS: Range verification was achieved by comparing the measured positron emission tomography activity distributions with the corresponding Monte Carlo-simulated distributions. Observed differences in the distal end of the activity distributions were analyzed as potential indicators for the range differences between the actual delivered and planned dose. RESULTS: The average spatial agreement between the measured and simulated activity distribution was within ±3 mm, and the corresponding average absolute agreement was within ±45% (derived from gamma index analysis). The mean absolute range deviation at 93 randomly chosen positions in 17 treatment fields delivered to 11 patients was 3.6 mm. Characteristic differences in the agreement of the measured and simulated activity distribution for the different tumor/target sites were found. This resulted from the different effect of factors such as biologic washout effects, motion, or limitations in the Monte Carlo-simulated activity patterns. CONCLUSION: We found that intracranial and cervical spine patients can greatly benefit from off-line positron emission tomography and computed tomography range verification. However, for the successful application of the method to patients with abdominopelvic tumors, major technological and methodologic improvements are needed. Among the intracranial and cervical spine target sites, patients with arteriovenous malformations or metal implants represent groups that could especially benefit from the approach.

MLC leaf width impact on the clinical dose distribution: a Monte Carlo approach.

PURPOSE: The influence of the multileaf collimator (MLC) leaf width on the dose distribution in patients treated with conformal radiotherapy and intensity-modulated radiotherapy has been analyzed. This study was based on the Monte Carlo simulation with the beams generated by a linac with the double-focused MLC. MATERIALS AND METHODS: The transmission through the leaves and the exact shape of the penumbra regions are difficult to model by treatment planning system algorithms. An accurate assessment of the dose variations due to the leaf width change can be achieved by means of Monte Carlo simulation. The BEAM/EGS4 code was used at the Hospital of the Virgen Macarena to model a Siemens PRIMUS linac, featuring an MLC with a leaf width projecting 1 cm at the isocenter. Based on this real model, a virtual head was designed while allowing for a variation of the leaf width projection. Both the real linac and the virtual linac, with leaves projecting 0.5 cm, were used to obtain the dose distributions for several treatments. A few disease sites, including the prostate, head and neck, and endometrium, were selected for the design of the conformal and intensity-modulated radiotherapy treatments with a forward planning algorithm sensitive to the different shapes of the volumes of interest. Isodose curves, differential matrix, gamma function, and the dose-volume histograms (DVHs) corresponding to both MLC models were obtained for all cases. The tumor control probability and the normal tissue complication probability were derived for those cases studied featuring the greatest differences between results for both MLCs. RESULTS: The impact on the DVHs of changing leaf width projections at the isocenter from 1.0 cm to 0.5 cm was low. Radiobiologic models showed slightly better tumor control probability/normal tissue complication probability values using the virtual MLC with a leaf width projecting 0.5 cm at isocenter in those cases presenting greater differences in the DVHs. CONCLUSIONS: The impact on the clinical dose distribution due to the MLC leaf width change is low based on the design and conditions used in this study.

The scored Patient-generated Subjective Global Assessment (PG-SGA) and its association with quality of life in ambulatory patients receiving radiotherapy.

OBJECTIVE: To evaluate the scored Patient-generated Subjective Global Assessment (PG-SGA) tool as an outcome measure in clinical nutrition practice and determine its association with quality of life (QoL). DESIGN: A prospective 4 week study assessing the nutritional status and QoL of ambulatory patients receiving radiation therapy to the head, neck, rectal or abdominal area. SETTING: Australian radiation oncology facilities. SUBJECTS: Sixty cancer patients aged 24-85 y. INTERVENTION: Scored PG-SGA questionnaire, subjective global assessment (SGA), QoL (EORTC QLQ-C30 version 3). RESULTS: According to SGA, 65.0% (39) of subjects were well-nourished, 28.3% (17) moderately or suspected of being malnourished and 6.7% (4) severely malnourished. PG-SGA score and global QoL were correlated (r=-0.66, P<0.001) at baseline. There was a decrease in nutritional status according to PG-SGA score (P<0.001) and SGA (P<0.001); and a decrease in global QoL (P<0.001) after 4 weeks of radiotherapy. There was a linear trend for change in PG-SGA score (P<0.001) and change in global QoL (P=0.003) between those patients who improved (5%) maintained (56.7%) or deteriorated (33.3%) in nutritional status according to SGA. There was a correlation between change in PG-SGA score and change in QoL after 4 weeks of radiotherapy (r=-0.55, P<0.001). Regression analysis determined that 26% of the variation of change in QoL was explained by change in PG-SGA (P=0.001). CONCLUSION: The scored PG-SGA is a nutrition assessment tool that identifies malnutrition in ambulatory oncology patients receiving radiotherapy and can be used to predict the magnitude of change in QoL.

Using knowledge maintenance for preference assessment.

Most real-life decisions require the decision maker to make trade-offs in order to fulfill multiple conflicting objectives. This is especially true in medical decision making while selecting the optimal therapy plan from among competing therapy plans for a patient. Multi-attribute utility theory provides a framework to specify these trade-offs for optimal decision making based on the preferences of the decision maker. However traditional preference-assessment techniques are difficult to implement and rarely elicit the true preferences of the decision maker. We describe a new preference-assessment method based on the concept of knowledge maintenance where the preference model is changed each time it makes an incorrect recommendation. The method is implemented in a decision-theoretic system to evaluate competing three-dimensional radiation treatment plans. The preference-assessment method leads to preference models which perform better than preference models elicited using traditional assessment techniques.

Assessment of renal function following irradiation of the intact spleen for Hodgkin disease.

Twenty-three patients with clinical Stage IA-IIIB Hodgkin disease underwent extended-field radiotherapy, including the intact spleen. In 17 of those patients, there was little evidence of renal dysfunction resulting from partial irradiation of the left kidney, which is inherent in such treatment. While isotopic images revealed an anatomic defect in the upper pole of the left kidney in one-third of the patients, dynamic studies showed that this anatomic alteration was not accompanied by any demonstrable dysfunction. Consistently normal serum BUN and creatinine levels in all patients tended to affirm that conclusion.

The role of ultrasound in the assessment and treatment of abdominal metastases from testicular tumours.

The para-aortic region has been investigated with ultrasound in 95 patients with testicular teratoma seen in the testicular tumour unit between 1973 and 1976. The liver was assessed in 110 patients. The technique has provided useful and reliable information. Para-aortic ultrasound can provide more information than existing diagnostic methods and is particularly useful for defining the volume of metastatic nodal masses. Its role should be regarded as complementary to that of lymphography. Ultrasonography is particularly useful to the radiotherapist since it can define the volume and limits of the tumour and thus influence choice of therapeutic approach and radiation planning technique. Ultrasound is also valuable in the assessment of response to treatment and in monitoring the patient during the subsequent follow-up period.