The application of artificial intelligence in the detection of basal cell carcinoma: A systematic review.

Basal cell carcinoma (BCC) is one of the most common types of cancer. The growing incidence worldwide and the need for fast, reliable and less invasive diagnostic techniques make a strong case for the application of different artificial intelligence techniques for detecting and classifying BCC and its subtypes. We report on the current evidence regarding the application of handcrafted and deep radiomics models used for the detection and classification of BCC in dermoscopy, optical coherence tomography and reflectance confocal microscopy. We reviewed all the articles that were published in the last 10 years in PubMed, Web of Science and EMBASE, and we found 15 articles that met the inclusion criteria. We included articles that are original, written in English, focussing on automated BCC detection in our target modalities and published within the last 10 years in the field of dermatology. The outcomes from the selected publications are presented in three categories depending on the imaging modality and to allow for comparison. The majority of articles (n = 12) presented different AI solutions for the detection and/or classification of BCC in dermoscopy images. The rest of the publications presented AI solutions in OCT images (n = 2) and RCM (n = 1). In addition, we provide future directions for the application of these techniques for the detection of BCC. In conclusion, the reviewed publications demonstrate the potential benefit of AI in the detection of BCC in dermoscopy, OCT and RCM.

Test-Retest Data for the Assessment of Breast MRI Radiomic Feature Repeatability.

BACKGROUND: Radiomic features extracted from breast MRI have potential for diagnostic, prognostic, and predictive purposes. However, before they can be used as biomarkers in clinical decision support systems, features need to be repeatable and reproducible. OBJECTIVE: Identify repeatable radiomics features within breast tissue on prospectively collected MRI exams through multiple test-retest measurements. STUDY TYPE: Prospective. POPULATION: 11 healthy female volunteers. FIELD STRENGTH/SEQUENCE: 1.5 T; MRI exams, comprising T2-weighted turbo spin-echo (T2W) sequence, native T1-weighted turbo gradient-echo (T1W) sequence, diffusion-weighted imaging (DWI) sequence using b-values 0/150/800, and corresponding derived ADC maps. ASSESSMENT: 18 MRI exams (three test-retest settings, repeated on 2 days) per healthy volunteer were examined on an identical scanner using a fixed clinical breast protocol. For each scan, 91 features were extracted from the 3D manually segmented right breast using Pyradiomics, before and after image preprocessing. Image preprocessing consisted of 1) bias field correction (BFC); 2) z-score normalization with and without BFC; 3) grayscale discretization using 32 and 64 bins with and without BFC; and 4) z-score normalization + grayscale discretization using 32 and 64 bins with and without BFC. STATISTICAL TESTS: Features' repeatability was assessed using concordance correlation coefficient(CCC) for each pair, i.e. each MRI was compared to each of the remaining 17 MRI with a cut-off value of CCC > 0.90. RESULTS: Images without preprocessing produced the highest number of repeatable features for both T1W sequence and ADC maps with 15 of 91 (16.5%) and 8 of 91 (8.8%) repeatable features, respectively. Preprocessed images produced between 4 of 91 (4.4%) and 14 of 91 (15.4%), and 6 of 91 (6.6%) and 7 of 91 (7.7%) repeatable features, respectively for T1W and ADC maps. Z-score normalization produced highest number of repeatable features, 26 of 91 (28.6%) in T2W sequences, in these images, no preprocessing produced 11 of 91 (12.1%) repeatable features. DATA CONCLUSION: Radiomic features extracted from T1W, T2W sequences and ADC maps from breast MRI exams showed a varying number of repeatable features, depending on the sequence. Effects of different preprocessing procedures on repeatability of features were different for each sequence. LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 1.

Radiomics for precision medicine: Current challenges, future prospects, and the proposal of a new framework.

The advancement of artificial intelligence concurrent with the development of medical imaging techniques provided a unique opportunity to turn medical imaging from mostly qualitative, to further quantitative and mineable data that can be explored for the development of clinical decision support systems (cDSS). Radiomics, a method for the high throughput extraction of hand-crafted features from medical images, and deep learning -the data driven modeling techniques based on the principles of simplified brain neuron interactions, are the most researched quantitative imaging techniques. Many studies reported on the potential of such techniques in the context of cDSS. Such techniques could be highly appealing due to the reuse of existing data, automation of clinical workflows, minimal invasiveness, three-dimensional volumetric characterization, and the promise of high accuracy and reproducibility of results and cost-effectiveness. Nevertheless, there are several challenges that quantitative imaging techniques face, and need to be addressed before the translation to clinical use. These challenges include, but are not limited to, the explainability of the models, the reproducibility of the quantitative imaging features, and their sensitivity to variations in image acquisition and reconstruction parameters. In this narrative review, we report on the status of quantitative medical image analysis using radiomics and deep learning, the challenges the field is facing, propose a framework for robust radiomics analysis, and discuss future prospects.

Dynamics of rectal balloon implant shrinkage in prostate VMAT : Influence on anorectal dose and late rectal complication risk.

PURPOSE: To assess the effect of a shrinking rectal balloon implant (RBI) on the anorectal dose and complication risk during the course of moderately hypofractionated prostate radiotherapy. METHODS: In 15 patients with localized prostate cancer, an RBI was implanted. A weekly kilovolt cone-beam computed tomography (CBCT) scan was acquired to measure the dynamics of RBI volume and prostate-rectum separation. The absolute anorectal volume encompassed by the 2 Gy equieffective 75 Gy isodose (V75Gy) was recalculated as well as the mean anorectal dose. The increase in estimated risk of grade 2-3 late rectal bleeding (LRB) between the start and end of treatment was predicted using nomograms. The observed acute and late toxicities were evaluated. RESULTS: A significant shrinkage of RBI volumes was observed, with an average volume of 70.4% of baseline at the end of the treatment. Although the prostate-rectum separation significantly decreased over time, it remained at least 1 cm. No significant increase in V75Gy of the anorectum was observed, except in one patient whose RBI had completely deflated in the third week of treatment. No correlation between mean anorectal dose and balloon deflation was found. The increase in predicted LRB risk was not significant, except in the one patient whose RBI completely deflated. The observed toxicities confirmed these findings. CONCLUSIONS: Despite significant decrease in RBI volume the high-dose rectal volume and the predicted LRB risk were unaffected due to a persistent spacing between the prostate and the anterior rectal wall.

PRODIGE: PRediction models in prOstate cancer for personalized meDIcine challenGE.

AIM: Identifying the best care for a patient can be extremely challenging. To support the creation of multifactorial Decision Support Systems (DSSs), we propose an Umbrella Protocol, focusing on prostate cancer. MATERIALS & METHODS: The PRODIGE project consisted of a workflow for standardizing data, and procedures, to create a consistent dataset useful to elaborate DSSs. Techniques from classical statistics and machine learning will be adopted. The general protocol accepted by our Ethical Committee can be downloaded from cancerdata.org . RESULTS: A standardized knowledge sharing process has been implemented by using a semi-formal ontology for the representation of relevant clinical variables. CONCLUSION: The development of DSSs, based on standardized knowledge, could be a tool to achieve a personalized decision-making.

Different prognostic models for different patient populations: validation of a new prognostic model for patients with oropharyngeal cancer in Western Europe.

OBJECTIVE: The presence of human papillomavirus (HPV) infection in oropharyngeal squamous cell carcinoma (OPSCC) is a major determinant in prognostic risk modelling. Recently, a prognostic model was proposed in which HPV status, comorbidity and nodal stage were the most important prognostic factors to determine high-, intermediate- and low-risk survival groups. Here, we report on the validation of this model using an independent single-institutional cohort. METHODS: A total number of 235 patients curatively treated for OPSCC in the period 2000-2011 at the MUMC (Maastricht University Medical Center, The Netherlands) were included. The presence of an oncogenic HPV infection was determined by p16 immunostaining, followed by a high-risk HPV DNA PCR on the p16-positive cases. The model variables included were HPV status, comorbidity and nodal stage. As a measure of model performance, the Harrell's Concordance index (Harrell's C-index) was used. RESULTS: The 5-year overall survival (OS) estimates were 84.6%, 54.5% and 28.7% in the low-, intermediate- and high-risk group, respectively. The difference between the survival curves was highly significant (P<0.001). The Harrell's C-index was 0.69 (95% confidence interval (CI): 0.63-0.75). CONCLUSION: In this study a previously developed prognostic risk model was validated. This model will help to personalise treatment in OPSCC patients. This model is publicly available at www.predictcancer.org.

Nomograms to predict survival and the risk for developing local or distant recurrence in patients with rectal cancer treated with optional short-term radiotherapy.

BACKGROUND: In many European countries, short-term 5 × 5 Gy radiotherapy has become the standard preoperative treatment of patients with resectable rectal cancer. Individualized risk assessment might allow a better selection of patients who will benefit from postoperative treatment and intensified follow-up. PATIENTS AND METHODS: From patient's data from three European rectal cancer trials (N = 2881), we developed multivariate cox nomograms reflecting the risk for local recurrence (LR), distant metastases (DM) and overall survival (OS). Evaluated variables were age, gender, tumour distance from the anal verge, the use of radiotherapy, surgical technique (total mesorectal excision/conventional surgery), surgery type (low anterior resection/abdominoperineal resection), time from randomization to surgery, residual disease (R0 versus R1 + 2), pT-stage, pN-stage and surgical complications. RESULTS: Pathological T- and N-status are of vital importance for an accurate prediction of LR, DM and OS. Short-course radiotherapy reduces the rate of LR. The developed nomograms are capable of predicting events with a validation c-index of 0.79 (LR), 0.76 (DM) and 0.75 (OS). The proposed stratification in risk groups allowed significant distinction between Kaplan-Meier curves for outcome. CONCLUSION: The developed nomograms can contribute to better individual risk prediction for LR, DM and OS for patients operated on rectal cancer. The practicality of the defined risk groups makes decision support in the consulting room feasible, assisting physicians to select patients for adjuvant therapy or intensified follow-up.

Intratumor genetic heterogeneity and head and neck cancer relapse.

BACKGROUND: Head and neck squamous cell carcinomas are treated by surgery, radiotherapy (RT), chemoradiotherapy (CRT) or combinations thereof, but locoregional recurrences (LRs) occur in 30-40% of treated patients. We have previously shown that in approximately half of the LRs after CRT, cancer driver mutations are not shared with the index tumor. AIM: To investigate two possible explanations for these genetically unrelated relapses, treatment-induced genetic changes and intratumor genetic heterogeneity. METHODS: To investigate treatment-induced clonal DNA changes, we compared copy number alterations (CNAs) and mutations between primary and recurrent xenografted tumors after treatment with (C)RT. Intratumor genetic heterogeneity was studied by multi-region sequencing on DNA from 31 biopsies of 11 surgically treated tumors. RESULTS: Induction of clonal DNA changes by (C)RT was not observed in the xenograft models. Multi-region sequencing demonstrated variations in CNA profiles between paired biopsies of individual tumors, with copy number heterogeneity scores varying from 0.027 to 0.333. In total, 32 cancer driver mutations could be identified and were shared in all biopsies of each tumor. Remarkably, multi-clonal mutations in these same cancer driver genes were observed in 6 of 11 tumors. Genetically distinct heterogeneous cell cultures could also be established from single tumors, with different biomarker profiles and drug sensitivities. CONCLUSION: Intratumor genetic heterogeneity at the level of the cancer driver mutations might explain the discordant mutational profiles in LRs after CRT, while there are no indications in xenograft models that these changes are induced by CRT.

Towards texture accurate slice interpolation of medical images using PixelMiner.

Quantitative image analysis models are used for medical imaging tasks such as registration, classification, object detection, and segmentation. For these models to be capable of making accurate predictions, they need valid and precise information. We propose PixelMiner, a convolution-based deep-learning model for interpolating computed tomography (CT) imaging slices. PixelMiner was designed to produce texture-accurate slice interpolations by trading off pixel accuracy for texture accuracy. PixelMiner was trained on a dataset of 7829 CT scans and validated using an external dataset. We demonstrated the model's effectiveness by using the structural similarity index (SSIM), peak signal to noise ratio (PSNR), and the root mean squared error (RMSE) of extracted texture features. Additionally, we developed and used a new metric, the mean squared mapped feature error (MSMFE). The performance of PixelMiner was compared to four other interpolation methods: (tri-)linear, (tri-)cubic, windowed sinc (WS), and nearest neighbor (NN). PixelMiner produced texture with a significantly lowest average texture error compared to all other methods with a normalized root mean squared error (NRMSE) of 0.11 (p < .01), and the significantly highest reproducibility with a concordance correlation coefficient (CCC) ≥ 0.85 (p < .01). PixelMiner was not only shown to better preserve features but was also validated using an ablation study by removing auto-regression from the model and was shown to improve segmentations on interpolated slices.

Independent and functional validation of a multi-tumour-type proliferation signature.

BACKGROUND: Previously we demonstrated that an mRNA signature reflecting cellular proliferation had strong prognostic value. As clinical applicability of signatures can be controversial, we sought to improve our marker's clinical utility by validating its biological relevance, reproducibility in independent data sets and applicability using an independent technique. METHODS: To facilitate signature evaluation with quantitative PCR (qPCR) a novel computational procedure was used to reduce the number of signature genes without significant information loss. These genes were validated in different human cancer cell lines upon serum starvation and in a 168 xenografts panel. Analyses were then extended to breast cancer and non-small-cell lung cancer (NSCLC) patient cohorts. RESULTS: Expression of the qPCR-based signature was dramatically decreased under starvation conditions and inversely correlated with tumour volume doubling time in xenografts. The signature validated in breast cancer (hazard ratio (HR)=1.63, P<0.001, n=1820) and NSCLC adenocarcinoma (HR=1.64, P<0.001, n=639) microarray data sets. Lastly, qPCR in a node-negative, non-adjuvantly treated breast cancer cohort (n=129) showed that patients assigned to the high-proliferation group had worse disease-free survival (HR=2.25, P<0.05). CONCLUSION: We have developed and validated a qPCR-based proliferation signature. This test might be used in the clinic to select (early-stage) patients for specific treatments that target proliferation.

Relocatable fixation systems in intracranial stereotactic radiotherapy. Accuracy of serial CT scans and patient acceptance in a randomized design.

PURPOSE: The goal was to provide a quantitative evaluation of the accuracy of three different fixation systems for stereotactic radiotherapy and to evaluate patients' acceptance for all fixations. METHODS: A total of 16 consecutive patients with brain tumours undergoing fractionated stereotactic radiotherapy (SCRT) were enrolled after informed consent (Clinical trials.gov: NCT00181350). Fixation systems evaluated were the BrainLAB® mask, with and without custom made bite-block (fixations S and A) and a homemade neck support with bite-block (fixation B) based on the BrainLAB® frame. The sequence of measurements was evaluated in a randomized manner with a cross-over design and patients' acceptance by a questionnaire. RESULTS: The mean three-dimensional (3D) displacement and standard deviations were 1.16 ± 0.68 mm for fixation S, 1.92 ± 1.28 and 1.70 ± 0.83 mm for fixations A and B, respectively. There was a significant improvement of the overall alignment (3D vector) when using the standard fixation instead of fixation A or B in the craniocaudal direction (p = 0.037). Rotational deviations were significantly less for the standard fixation S in relation to fixations A (p = 0.005) and B (p = 0.03). EPI imaging with off-line correction further improved reproducibility. Five out of 8 patients preferred the neck support with the bite-block. CONCLUSION: The mask fixation system in conjunction with a bite-block is the most accurate fixation for SCRT reducing craniocaudal and rotational movements. Patients favoured the more comfortable but less accurate neck support. To optimize the accuracy of SCRT, additional regular portal imaging is warranted.

Radiation-induced oesophagitis in lung cancer patients. Is susceptibility for neutropenia a risk factor?

BACKGROUND: Radiation-induced oesophagitis is a major side effect of concurrent chemotherapy and radiotherapy. A strong association between neutropenia and oesophagitis was previously shown, but external validation and further elucidation of the possible mechanisms are lacking. METHODS AND PATIENTS: A total of 119 patients were included at two institutions. The concurrent group comprised 34 SCLC patients treated with concurrent carboplatin and etoposide, and concurrent chest irradiation, and 36 NSCLC patients with concurrent cisplatin and etoposide, and concurrent radiotherapy, while the sequential group comprised 49 NSCLC patients received sequential cisplatin and gemcitabine, and radiotherapy. RESULTS: Severe neutropenia was very frequent during concurrent chemoradiation (grade: 4 41.4%) and during induction chemotherapy in sequentially treated patients (grade 4: 30.6%), but not during radiotherapy (only 4% grade 1). In the concurrent group, the odds ratios of grade 3 oesophagitis vs. neutropenia were the following: grade 2 vs. grade 0/1: 5.60 (95% CI 1.55-20.26), p = 0.009; grade 3 vs. grade 0/1: 10.40 (95% CI 3.19-33.95); p = 0.0001; grade 4 vs. grade 0/1: 12.60 (95% CI 4.36-36.43); p < 0.00001. There was no correlation between the occurrence of neutropenia during induction chemotherapy and acute oesophagitis during or after radiotherapy alone. In the univariate analysis, total radiation dose (p < 0.001), overall treatment time of radiotherapy (p < 0.001), mean oesophageal dose (p = 0.038) and neutropenia (p < 0.001) were significantly associated with the development of oesophagitis. In a multivariate analysis, only neutropenia remained significant (p = 0.023). CONCLUSION: We confirm that neutropenia is independently correlated with oesophagitis in concurrent chemoradiation, but that the susceptibility for chemotherapy-induced neutropenia is not associated with radiation-induced oesophagitis. Further studies focusing on the underlying mechanisms are thus warranted.

Lhermitte sign and myelopathy after irradiation of the cervical spinal cord in radiotherapy treatment of head and neck cancer.

BACKGROUND AND PURPOSE: The goal of this work was to examine toxicity and risk factors after irradiation of the cervical spinal cord. PATIENTS AND METHODS: A total of 437 patients irradiated for a laryngeal and oropharyngeal carcinoma were eligible (median follow-up 27 months). Spinal cord contouring was defined differently over time as anatomically defined spinal cord area (SCA) and the spinal cord on CT (SC) with a margin of 3 or 5 mm (SCP3/SCP5). RESULTS: None developed chronic progressive radiation myelopathy (CPRM) (maximum spinal dose 21.8-69 Gy); 3.9% (17/437) developed a Lhermitte sign (LS) with a median duration of 6 months (range 1-30 months) and was reversible in all patients. Risk factors for developing LS were younger age (52 vs. 61 years, p < 0.001), accelerated RT (12/17 patients, p < 0.005), and dose-volume relationships for SCA with ≥ 45 Gy of 14.15 cm(3) and 7.9 cm(3) for patients with and without LS, respectively. CONCLUSION: LS is more frequently observed in younger patients and in patients treated with accelerated radiotherapy. A dose-volume relationship was seen for V45 in the case of SCA. For higher doses, no clear dose-volume relationships were observed.

Radiomics for Predicting Lung Cancer Outcomes Following Radiotherapy: A Systematic Review.

Lung cancer's radiomic phenotype may potentially inform clinical decision-making with respect to radical radiotherapy. At present there are no validated biomarkers available for the individualisation of radical radiotherapy in lung cancer and the mortality rate of this disease remains the highest of all other solid tumours. MEDLINE was searched using the terms 'radiomics' and 'lung cancer' according to the Preferred Reporting Items for Systematic Reviews and Met-Analyses (PRISMA) guidance. Radiomics studies were defined as those manuscripts describing the extraction and analysis of at least 10 quantifiable imaging features. Only those studies assessing disease control, survival or toxicity outcomes for patients with lung cancer following radical radiotherapy ± chemotherapy were included. Study titles and abstracts were reviewed by two independent reviewers. The Radiomics Quality Score was applied to the full text of included papers. Of 244 returned results, 44 studies met the eligibility criteria for inclusion. End points frequently reported were local (17%), regional (17%) and distant control (31%), overall survival (79%) and pulmonary toxicity (4%). Imaging features strongly associated with clinical outcomes include texture features belonging to the subclasses Gray level run length matrix, Gray level co-occurrence matrix and kurtosis. The median cohort size for model development was 100 (15-645); in the 11 studies with external validation in a separate independent population, the median cohort size was 84 (21-295). The median number of imaging features extracted was 184 (10-6538). The median Radiomics Quality Score was 11% (0-47). Patient-reported outcomes were not incorporated within any studies identified. No studies externally validated a radiomics signature in a registered prospective study. Imaging-derived indices attained through radiomic analyses could equip thoracic oncologists with biomarkers for treatment response, patterns of failure, normal tissue toxicity and survival in lung cancer. Based on routine scans, their non-invasive nature and cost-effectiveness are major advantages over conventional pathological assessment. Improved tools are required for the appraisal of radiomics studies, as significant barriers to clinical implementation remain, such as standardisation of input scan data, quality of reporting and external validation of signatures in randomised, interventional clinical trials.

Comparison of Bayesian network and support vector machine models for two-year survival prediction in lung cancer patients treated with radiotherapy.

PURPOSE: Classic statistical and machine learning models such as support vector machines (SVMs) can be used to predict cancer outcome, but often only perform well if all the input variables are known, which is unlikely in the medical domain. Bayesian network (BN) models have a natural ability to reason under uncertainty and might handle missing data better. In this study, the authors hypothesize that a BN model can predict two-year survival in non-small cell lung cancer (NSCLC) patients as accurately as SVM, but will predict survival more accurately when data are missing. METHODS: A BN and SVM model were trained on 322 inoperable NSCLC patients treated with radiotherapy from Maastricht and validated in three independent data sets of 35, 47, and 33 patients from Ghent, Leuven, and Toronto. Missing variables occurred in the data set with only 37, 28, and 24 patients having a complete data set. RESULTS: The BN model structure and parameter learning identified gross tumor volume size, performance status, and number of positive lymph nodes on a PET as prognostic factors for two-year survival. When validated in the full validation set of Ghent, Leuven, and Toronto, the BN model had an AUC of 0.77, 0.72, and 0.70, respectively. A SVM model based on the same variables had an overall worse performance (AUC 0.71, 0.68, and 0.69) especially in the Ghent set, which had the highest percentage of missing the important GTV size data. When only patients with complete data sets were considered, the BN and SVM model performed more alike. CONCLUSIONS: Within the limitations of this study, the hypothesis is supported that BN models are better at handling missing data than SVM models and are therefore more suitable for the medical domain. Future works have to focus on improving the BN performance by including more patients, more variables, and more diversity.

Exploring imaging features of molecular subtypes of large cell neuroendocrine carcinoma (LCNEC).

OBJECTIVES: Radiological characteristics and radiomics signatures can aid in differentiation between small cell lung carcinoma (SCLC) and non-small cell lung carcinoma (NSCLC). We investigated whether molecular subtypes of large cell neuroendocrine carcinoma (LCNEC), i.e. SCLC-like (with pRb loss) vs. NSCLC-like (with pRb expression), can be distinguished by imaging based on (1) imaging interpretation, (2) semantic features, and/or (3) a radiomics signature, designed to differentiate between SCLC and NSCLC. MATERIALS AND METHODS: Pulmonary oncologists and chest radiologists assessed chest CT-scans of 44 LCNEC patients for 'small cell-like' or 'non-small cell-like' appearance. The radiologists also scored semantic features of 50 LCNEC scans. Finally, a radiomics signature was trained on a dataset containing 48 SCLC and 76 NSCLC scans and validated on an external set of 58 SCLC and 40 NSCLC scans. This signature was applied on scans of 28 SCLC-like and 8 NSCLC-like LCNEC patients. RESULTS: Pulmonary oncologists and radiologists were unable to differentiate between molecular subtypes of LCNEC and no significant differences in semantic features were found. The area under the receiver operating characteristics curve of the radiomics signature in the validation set (SCLC vs. NSCLC) was 0.84 (95% confidence interval (CI) 0.77-0.92) and 0.58 (95% CI 0.29-0.86) in the LCNEC dataset (SCLC-like vs. NSCLC-like). CONCLUSION: LCNEC appears to have radiological characteristics of both SCLC and NSCLC, irrespective of pRb loss, compatible with the SCLC-like subtype. Imaging interpretation, semantic features and our radiomics signature designed to differentiate between SCLC and NSCLC were unable to separate molecular LCNEC subtypes, which underscores that LCNEC is a unique disease.

Eligibility for concurrent chemotherapy and radiotherapy of locally advanced lung cancer patients: a prospective, population-based study.

BACKGROUND: Patients with stage III non-small-cell lung cancer (NSCLC) and limited disease small-cell lung cancer are excluded from concurrent chemoradiation mostly on the basis of comorbidity and age. The purpose of this prospective study was to get insight in what proportion of patients with locally advanced lung cancer would be suitable for concurrent chemoradiation. PATIENTS AND METHODS: From 2002 to 2005, all patients with a pathological diagnosis of lung cancer and with locally advanced disease in the Maastricht Cancer Registry, the Netherlands, comorbidity were prospectively assessed. Patients were regarded as noneligible for concurrent chemoradiation if they had one or more important comorbidity or were 75 years or older. RESULTS: In all, 711 patients were included, 577 with NSCLC and 134 with SCLC. Overall, 166 patients (23.3%) were 75 years or older. Of the 526 patients <75 years, comorbidities were as follows: 278 (52.9%) 0, 188 (35.7%) 1, and 56 (11.4%) 2 or more. In all, 408/686 (59%) of the whole patient group were considered as ineligible for concurrent chemoradiation. CONCLUSIONS: More than half of patients with stage III lung cancer were theoretically not eligible for concurrent chemoradiation. Less toxic alternatives are needed for these patients.

Prediction of DVH parameter changes due to setup errors for breast cancer treatment based on 2D portal dosimetry.

Electronic portal imaging devices (EPIDs) are increasingly used for portal dosimetry applications. In our department, EPIDs are clinically used for two-dimensional (2D) transit dosimetry. Predicted and measured portal dose images are compared to detect dose delivery errors caused for instance by setup errors or organ motion. The aim of this work is to develop a model to predict dose-volume histogram (DVH) changes due to setup errors during breast cancer treatment using 2D transit dosimetry. First, correlations between DVH parameter changes and 2D gamma parameters are investigated for different simulated setup errors, which are described by a binomial logistic regression model. The model calculates the probability that a DVH parameter changes more than a specific tolerance level and uses several gamma evaluation parameters for the planning target volume (PTV) projection in the EPID plane as input. Second, the predictive model is applied to clinically measured portal images. Predicted DVH parameter changes are compared to calculated DVH parameter changes using the measured setup error resulting from a dosimetric registration procedure. Statistical accuracy is investigated by using receiver operating characteristic (ROC) curves and values for the area under the curve (AUC), sensitivity, specificity, positive and negative predictive values. Changes in the mean PTV dose larger than 5%, and changes in V90 and V95 larger than 10% are accurately predicted based on a set of 2D gamma parameters. Most pronounced changes in the three DVH parameters are found for setup errors in the lateral-medial direction. AUC, sensitivity, specificity, and negative predictive values were between 85% and 100% while the positive predictive values were lower but still higher than 54%. Clinical predictive value is decreased due to the occurrence of patient rotations or breast deformations during treatment, but the overall reliability of the predictive model remains high. Based on our predictive model, 2D transit dosimetry measurements can now directly be translated in clinically more relevant DVH parameter changes for the PTV during conventional breast treatment. In this way, the possibility to design decision protocols based on extracted DVH changes is created instead of undertaking elaborate actions such as repeated treatment planning or 3D dose reconstruction for a large group of patients.

Decision Support Systems in Prostate Cancer Treatment: An Overview.

BACKGROUND: A multifactorial decision support system (mDSS) is a tool designed to improve the clinical decision-making process, while using clinical inputs for an individual patient to generate case-specific advice. The study provides an overview of the literature to analyze current available mDSS focused on prostate cancer (PCa), in order to better understand the availability of decision support tools as well as where the current literature is lacking. METHODS: We performed a MEDLINE literature search in July 2018. We divided the included studies into different sections: diagnostic, which aids in detection or staging of PCa; treatment, supporting the decision between treatment modalities; and patient, which focusses on informing the patient. We manually screened and excluded studies that did not contain an mDSS concerning prostate cancer and study proposals. RESULTS: Our search resulted in twelve diagnostic mDSS; six treatment mDSS; two patient mDSS; and eight papers that could improve mDSS. CONCLUSIONS: Diagnosis mDSS is well represented in the literature as well as treatment mDSS considering external-beam radiotherapy; however, there is a lack of mDSS for other treatment modalities. The development of patient decision aids is a new field of research, and few successes have been made for PCa patients. These tools can improve personalized medicine but need to overcome a number of difficulties to be successful and require more research.