Endoscopic ultrasound for structured surveillance after curative treatment of esophageal cancer.

BACKGROUND: Structured surveillance after treatment of esophageal cancer is not established. Due to a paucity of data, no agreement exists on how surveillance should be performed. The main argument against intensive follow-up in esophageal cancer is that it may not lead to true survival advantage. METHODS: Structured surveillance was performed in 42 patients after multimodal therapy with peri-operative chemotherapy (29) or definitive chemoradiotherapy (13) of esophageal cancer. The surveillance protocol included gastroscopy, endoscopic ultrasound, chest X-ray, abdominal ultrasound, and CEA measurement at regular intervals of up to five years. We analyzed relapse rate, time to relapse, localization of recurrence, diagnosis within or without structured surveillance, diagnostic method providing the first evidence of a relapse, treatment of recurrence, and outcome. RESULTS: Median follow-up was 48 months; 18/42 patients suffered from tumor relapse, with 16 asymptomatic patients diagnosed within structured surveillance. Median time to recurrence was 9 months. Isolated local or locoregional recurrence occurred in 6, and isolated distant relapse in 9 patients. All patients with isolated locoregional recurrence were exclusively diagnosed with endoscopic ultrasound. Six patients received curatively intended therapy with surgery or chemoradiation, leading to long-lasting survival. CONCLUSION: Structured surveillance offers the chance to identify limited and asymptomatic tumor relapse. Especially in cases of locoregional recurrence, long-lasting survival or even a cure can be achieved. Endoscopic ultrasound is the best method for the detection of locoregional tumor recurrence and should be an integral part of structured surveillance after curative treatment of esophageal cancer.

Direct comparison of multiple computer-aided polyp detection systems.

BACKGROUND AND STUDY AIMS: Artificial intelligence (AI)-based systems for computer-aided detection (CADe) of polyps receive regular updates and occasionally offer customizable detection thresholds, both of which impact their performance, but little is known about these effects. This study aimed to compare the performance of different CADe systems on the same benchmark dataset. METHODS: 101 colonoscopy videos were used as benchmark. Each video frame with a visible polyp was manually annotated with bounding boxes, resulting in 129 705 polyp images. The videos were then analyzed by three different CADe systems, representing five conditions: two versions of GI Genius, Endo-AID with detection Types A and B, and EndoMind, a freely available system. Evaluation included an analysis of sensitivity and false-positive rate, among other metrics. RESULTS: Endo-AID detection Type A, the earlier version of GI Genius, and EndoMind detected all 93 polyps. Both the later version of GI Genius and Endo-AID Type B missed 1 polyp. The mean per-frame sensitivities were 50.63 % and 67.85 %, respectively, for the earlier and later versions of GI Genius, 65.60 % and 52.95 %, respectively, for Endo-AID Types A and B, and 60.22 % for EndoMind. CONCLUSIONS: This study compares the performance of different CADe systems, different updates, and different configuration modes. This might help clinicians to select the most appropriate system for their specific needs.

Assisted documentation as a new focus for artificial intelligence in endoscopy: the precedent of reliable withdrawal time and image reporting.

BACKGROUND : Reliable documentation is essential for maintaining quality standards in endoscopy; however, in clinical practice, report quality varies. We developed an artificial intelligence (AI)-based prototype for the measurement of withdrawal and intervention times, and automatic photodocumentation. METHOD: A multiclass deep learning algorithm distinguishing different endoscopic image content was trained with 10 557 images (1300 examinations, nine centers, four processors). Consecutively, the algorithm was used to calculate withdrawal time (AI prediction) and extract relevant images. Validation was performed on 100 colonoscopy videos (five centers). The reported and AI-predicted withdrawal times were compared with video-based measurement; photodocumentation was compared for documented polypectomies. RESULTS: Video-based measurement in 100 colonoscopies revealed a median absolute difference of 2.0 minutes between the measured and reported withdrawal times, compared with 0.4 minutes for AI predictions. The original photodocumentation represented the cecum in 88 examinations compared with 98/100 examinations for the AI-generated documentation. For 39/104 polypectomies, the examiners' photographs included the instrument, compared with 68 for the AI images. Lastly, we demonstrated real-time capability (10 colonoscopies). CONCLUSION : Our AI system calculates withdrawal time, provides an image report, and is real-time ready. After further validation, the system may improve standardized reporting, while decreasing the workload created by routine documentation.

Automated classification of polyps using deep learning architectures and few-shot learning.

BACKGROUND: Colorectal cancer is a leading cause of cancer-related deaths worldwide. The best method to prevent CRC is a colonoscopy. However, not all colon polyps have the risk of becoming cancerous. Therefore, polyps are classified using different classification systems. After the classification, further treatment and procedures are based on the classification of the polyp. Nevertheless, classification is not easy. Therefore, we suggest two novel automated classifications system assisting gastroenterologists in classifying polyps based on the NICE and Paris classification. METHODS: We build two classification systems. One is classifying polyps based on their shape (Paris). The other classifies polyps based on their texture and surface patterns (NICE). A two-step process for the Paris classification is introduced: First, detecting and cropping the polyp on the image, and secondly, classifying the polyp based on the cropped area with a transformer network. For the NICE classification, we design a few-shot learning algorithm based on the Deep Metric Learning approach. The algorithm creates an embedding space for polyps, which allows classification from a few examples to account for the data scarcity of NICE annotated images in our database. RESULTS: For the Paris classification, we achieve an accuracy of 89.35 %, surpassing all papers in the literature and establishing a new state-of-the-art and baseline accuracy for other publications on a public data set. For the NICE classification, we achieve a competitive accuracy of 81.13 % and demonstrate thereby the viability of the few-shot learning paradigm in polyp classification in data-scarce environments. Additionally, we show different ablations of the algorithms. Finally, we further elaborate on the explainability of the system by showing heat maps of the neural network explaining neural activations. CONCLUSION: Overall we introduce two polyp classification systems to assist gastroenterologists. We achieve state-of-the-art performance in the Paris classification and demonstrate the viability of the few-shot learning paradigm in the NICE classification, addressing the prevalent data scarcity issues faced in medical machine learning.

Artificial intelligence-based polyp size measurement in gastrointestinal endoscopy using the auxiliary waterjet as a reference.

BACKGROUND: Measurement of colorectal polyp size during endoscopy is mainly performed visually. In this work, we propose a novel polyp size measurement system (Poseidon) based on artificial intelligence (AI) using the auxiliary waterjet as a measurement reference. METHODS: Visual estimation, biopsy forceps-based estimation, and Poseidon were compared using a computed tomography colonography-based silicone model with 28 polyps of defined sizes. Four experienced gastroenterologists estimated polyp sizes visually and with biopsy forceps. Furthermore, the gastroenterologists recorded images of each polyp with the waterjet in proximity for the application of Poseidon. Additionally, Poseidon's measurements of 29 colorectal polyps during routine clinical practice were compared with visual estimates. RESULTS: In the silicone model, visual estimation had the largest median percentage error of 25.1 % (95 %CI 19.1 %-30.4 %), followed by biopsy forceps-based estimation: median 20.0 % (95 %CI 14.4 %-25.6 %). Poseidon gave a significantly lower median percentage error of 7.4 % (95 %CI 5.0 %-9.4 %) compared with other methods. During routine colonoscopies, Poseidon presented a significantly lower median percentage error (7.7 %, 95 %CI 6.1 %-9.3 %) than visual estimation (22.1 %, 95 %CI 15.1 %-26.9 %). CONCLUSION: In this work, we present a novel AI-based method for measuring colorectal polyp size with significantly higher accuracy than other common sizing methods.

A video based benchmark data set (ENDOTEST) to evaluate computer-aided polyp detection systems.

BACKGROUND AND AIMS: Computer-aided polyp detection (CADe) may become a standard for polyp detection during colonoscopy. Several systems are already commercially available. We report on a video-based benchmark technique for the first preclinical assessment of such systems before comparative randomized trials are to be undertaken. Additionally, we compare a commercially available CADe system with our newly developed one. METHODS: ENDOTEST consisted in the combination of two datasets. The validation dataset contained 48 video-snippets with 22,856 manually annotated images of which 53.2% contained polyps. The performance dataset contained 10 full-length screening colonoscopies with 230,898 manually annotated images of which 15.8% contained a polyp. Assessment parameters were accuracy for polyp detection and time delay to first polyp detection after polyp appearance (FDT). Two CADe systems were assessed: a commercial CADe system (GI-Genius, Medtronic), and a self-developed new system (ENDOMIND). The latter being a convolutional neuronal network trained on 194,983 manually labeled images extracted from colonoscopy videos recorded in mainly six different gastroenterologic practices. RESULTS: On the ENDOTEST, both CADe systems detected all polyps in at least one image. The per-frame sensitivity and specificity in full colonoscopies was 48.1% and 93.7%, respectively for GI-Genius; and 54% and 92.7%, respectively for ENDOMIND. Median FDT of ENDOMIND with 217 ms (Inter-Quartile Range(IQR)8-1533) was significantly faster than GI-Genius with 1050 ms (IQR 358-2767, p = 0.003). CONCLUSIONS: Our benchmark ENDOTEST may be helpful for preclinical testing of new CADe devices. There seems to be a correlation between a shorter FDT with a higher sensitivity and a lower specificity for polyp detection.

Pilot study of a new freely available computer-aided polyp detection system in clinical practice.

PURPOSE: Computer-aided polyp detection (CADe) systems for colonoscopy are already presented to increase adenoma detection rate (ADR) in randomized clinical trials. Those commercially available closed systems often do not allow for data collection and algorithm optimization, for example regarding the usage of different endoscopy processors. Here, we present the first clinical experiences of a, for research purposes publicly available, CADe system. METHODS: We developed an end-to-end data acquisition and polyp detection system named EndoMind. Examiners of four centers utilizing four different endoscopy processors used EndoMind during their clinical routine. Detected polyps, ADR, time to first detection of a polyp (TFD), and system usability were evaluated (NCT05006092). RESULTS: During 41 colonoscopies, EndoMind detected 29 of 29 adenomas in 66 of 66 polyps resulting in an ADR of 41.5%. Median TFD was 130 ms (95%-CI, 80-200 ms) while maintaining a median false positive rate of 2.2% (95%-CI, 1.7-2.8%). The four participating centers rated the system using the System Usability Scale with a median of 96.3 (95%-CI, 70-100). CONCLUSION: EndoMind's ability to acquire data, detect polyps in real-time, and high usability score indicate substantial practical value for research and clinical practice. Still, clinical benefit, measured by ADR, has to be determined in a prospective randomized controlled trial.

Fast machine learning annotation in the medical domain: a semi-automated video annotation tool for gastroenterologists.

BACKGROUND: Machine learning, especially deep learning, is becoming more and more relevant in research and development in the medical domain. For all the supervised deep learning applications, data is the most critical factor in securing successful implementation and sustaining the progress of the machine learning model. Especially gastroenterological data, which often involves endoscopic videos, are cumbersome to annotate. Domain experts are needed to interpret and annotate the videos. To support those domain experts, we generated a framework. With this framework, instead of annotating every frame in the video sequence, experts are just performing key annotations at the beginning and the end of sequences with pathologies, e.g., visible polyps. Subsequently, non-expert annotators supported by machine learning add the missing annotations for the frames in-between. METHODS: In our framework, an expert reviews the video and annotates a few video frames to verify the object's annotations for the non-expert. In a second step, a non-expert has visual confirmation of the given object and can annotate all following and preceding frames with AI assistance. After the expert has finished, relevant frames will be selected and passed on to an AI model. This information allows the AI model to detect and mark the desired object on all following and preceding frames with an annotation. Therefore, the non-expert can adjust and modify the AI predictions and export the results, which can then be used to train the AI model. RESULTS: Using this framework, we were able to reduce workload of domain experts on average by a factor of 20 on our data. This is primarily due to the structure of the framework, which is designed to minimize the workload of the domain expert. Pairing this framework with a state-of-the-art semi-automated AI model enhances the annotation speed further. Through a prospective study with 10 participants, we show that semi-automated annotation using our tool doubles the annotation speed of non-expert annotators compared to a well-known state-of-the-art annotation tool. CONCLUSION: In summary, we introduce a framework for fast expert annotation for gastroenterologists, which reduces the workload of the domain expert considerably while maintaining a very high annotation quality. The framework incorporates a semi-automated annotation system utilizing trained object detection models. The software and framework are open-source.

Development and evaluation of a deep learning model to improve the usability of polyp detection systems during interventions.

BACKGROUND: The efficiency of artificial intelligence as computer-aided detection (CADe) systems for colorectal polyps has been demonstrated in several randomized trials. However, CADe systems generate many distracting detections, especially during interventions such as polypectomies. Those distracting CADe detections are often induced by the introduction of snares or biopsy forceps as the systems have not been trained for such situations. In addition, there are a significant number of non-false but not relevant detections, since the polyp has already been previously detected. All these detections have the potential to disturb the examiner's work. OBJECTIVES: Development and evaluation of a convolutional neuronal network that recognizes instruments in the endoscopic image, suppresses distracting CADe detections, and reliably detects endoscopic interventions. METHODS: A total of 580 different examination videos from 9 different centers using 4 different processor types were screened for instruments and represented the training dataset (519,856 images in total, 144,217 contained a visible instrument). The test dataset included 10 full-colonoscopy videos that were analyzed for the recognition of visible instruments and detections by a commercially available CADe system (GI Genius, Medtronic). RESULTS: The test dataset contained 153,623 images, 8.84% of those presented visible instruments (12 interventions, 19 instruments used). The convolutional neuronal network reached an overall accuracy in the detection of visible instruments of 98.59%. Sensitivity and specificity were 98.55% and 98.92%, respectively. A mean of 462.8 frames containing distracting CADe detections per colonoscopy were avoided using the convolutional neuronal network. This accounted for 95.6% of all distracting CADe detections. CONCLUSIONS: Detection of endoscopic instruments in colonoscopy using artificial intelligence technology is reliable and achieves high sensitivity and specificity. Accordingly, the new convolutional neuronal network could be used to reduce distracting CADe detections during endoscopic procedures. Thus, our study demonstrates the great potential of artificial intelligence technology beyond mucosal assessment.

Treatment Intensity, not Age, Affects Survival Time of Patients with Advanced Pancreatic Cancer.

BACKGROUND AND AIMS: Despite older-aged individuals accounting for most patients with pancreatic cancer, elderly patients are still underrepresented in the clinical trials. Our study aims to identify treatment differences as well as to analyze survival times in the younger and older patient group. METHODS: We evaluated the data of 97 pancreatic cancer patients (72 <75 years; 25 ≥75 years) receiving palliative chemotherapy. Age, comorbidity, body mass index (BMI), tumor localization, metastases, carbohydrate-antigen 19-9 (CA19-9) value, number and type of chemotherapeutic agents and treatment regimens used, treatment lines, toxicity and survival time were assessed. RESULTS: The age groups did not differ in their initial conditions (comorbidity, BMI, tumor characteristics). However, treatment intensity of patients ≥ 75 years was lower. Elderly patients received significantly fewer different chemotherapeutic agents and therapeutic regimens, therapy lines and fewer combination chemotherapies. Moreover, elderly patients survived significantly shorter (7.6 vs. 12.7 months, p=0.001). In multivariance analysis, a significant negative influence on survival time was revealed for low therapy intensity (≤2 chemotherapeutics, ≤2 therapy lines), but not for age. In addition, therapy discontinuation and underweight were significantly associated with survival time. CONCLUSION: Not age per se but lower therapy intensity leads to a shorter overall survival in the elderly patient group.

Predictors of COVID-19 in an outpatient fever clinic.

BACKGROUND: The objective of this study was to identify clinical risk factors for COVID-19 in a German outpatient fever clinic that allow distinction of SARS-CoV-2 infected patients from other patients with flu-like symptoms. METHODS: This is a retrospective, single-centre cohort study. Patients were included visiting the fever clinic from 4th of April 2020 to 15th of May 2020. Symptoms, comorbidities, and socio-demographic factors were recorded in a standardized fashion. Multivariate logistic regression was used to identify risk factors of COVID-19, on the bases of those a model discrimination was assessed using area under the receiver operation curves (AUROC). RESULTS: The final analysis included 930 patients, of which 74 (8%) had COVID-19. Anosmia (OR 10.71; CI 6.07-18.9) and ageusia (OR 9.3; CI 5.36-16.12) were strongly associated with COVID-19. High-risk exposure (OR 12.20; CI 6.80-21.90), especially in the same household (OR 4.14; CI 1.28-13.33), was also correlated; the more household members, especially with flu-like symptoms, the higher the risk of COVID-19. Working in an essential workplace was also associated with COVID-19 (OR 2.35; CI 1.40-3.96), whereas smoking was inversely correlated (OR 0.19; CI 0.08-0.44). A model that considered risk factors like anosmia, ageusia, concomitant of symptomatic household members and smoking well discriminated COVID-19 patients from other patients with flu-like symptoms (AUROC 0.84). CONCLUSIONS: We report a set of four readily available clinical parameters that allow the identification of high-risk individuals of COVID-19. Our study will not replace molecular testing but will help guide containment efforts while waiting for test results.

Pooled SARS-CoV-2 antigen tests in asymptomatic children and their caregivers: Screening for SARS-CoV-2 in a pediatric emergency department.

BACKGROUND: Universal admission screening for SARS-CoV-2 in children and their caregivers (CG) is critical to prevent hospital outbreaks. We evaluated pooled SARS-CoV-2 antigen tests (AG) to identify infectious individuals while waiting for polymerase chain reaction (PCR) test results. METHODS: This single-center study was performed from November 5, 2020 to March 1, 2021. Nasal mid-turbinate and oropharyngeal swabbing for AG and PCR testing was performed in children with 2 individual swabs that were simultaneously inserted. Nasopharyngeal swabs were obtained from their CG. AG swabs were pooled in a single extraction buffer tube and PCR swabs in a single viral medium. Results from an adult population were used for comparison, as no pooled testing was performed. RESULTS: During the study period, 710 asymptomatic children and their CG were admitted. Pooled AG sensitivity and specificity was 75% and 99.4% respectively for detection of infectious individuals. Four false negatives were observed, though 3 out of 4 false negative child-CG pairs were not considered infectious at admission. Unpooled AG testing in an adult population showed a comparable sensitivity and specificity of 50% and 99.7%. AG performed significantly better in samples with lower Ct values in the corresponding PCR (32.3 vs 21, P-value < .001). CONCLUSIONS: Pooled SARS-CoV-2 AGs are an effective method to identify potentially contagious individuals prior admission, without adding additional strain to the child.

Predictors for Treatment Failure of Self-Expandable Metal Stents for Anastomotic Leak after Gastro-Esophageal Resection.

BACKGROUND AND AIMS: Self-expandable metal stents are used for the treatment of anastomotic leaks after gastro- esophageal surgery. Predictors for treatment failure and complications are unknown. In this observational retrospective study, we summarize our experience with self-expandable metal stents for the treatment of anastomotic leaks, in order to determine the predictors of treatment failure. METHODS: Between 2009 and 2015, 34 patients with anastomotic leak after curative resection of gastro- esophageal cancer were treated with self-expandable metal stents. Gender, histology, comorbidity, body mass index, neoadjuvant therapy, previous surgery, leak size, and stent diameter were analyzed for their predictive value according to treatment success and complication rate. RESULTS: Leak closure rate was 76%. Risk factors for treatment failure were neoadjuvant chemo-radiotherapy, squamous cell histology, and esophageal tumor location. Gender, comorbidity, body mass index, neoadjuvant chemotherapy, and previous surgery were not correlated with outcome. Mortality rate was 20%, most often due to uncontrolled leak. Severe stent-related complications occurred in 15% of patients, most of them following insertion of a large-sized stent. CONCLUSION: Squamous cell histology, neoadjuvant chemo-radiotherapy, and esophageal tumor location are predictors for treatment failure. Severe stent-related complications seem to be preferentially associated with the use of large-sized stents.

Vocal Fold Paralysis is a Common Complication in Esophageal Cancer Patients with Esophagorespiratory Fistula.

BACKGROUND AND AIMS: The development of an esophagorespiratory fistula (ERF) in patients with esophageal cancer (EC) is associated with poor prognosis. We observed a high rate of vocal fold paralysis (VFP) in patients with ERF. Data on prevalence and complications of VFP in ERF are lacking. The present study investigated the incidence of VFP in patients with malignant ERF and examined possible risk factors and the impact on survival. METHODS: We performed a retrospective case-control study of 46 institutional cases of EC patients with ERF in a time period of eleven years. Patients were matched to 92 randomly selected controls (EC patients without ERF) in a 1:2 fashion for tumor localization and histology. Demographics, clinical characteristics, recurrence, treatment modalities as well as survival were analyzed. RESULTS: Esophageal cancer patients with ERF developed more often VFP than EC patients without ERF (59% vs. 21%; p=0.02; odds ratio (OR) 4.9). Esophageal cancer patients with ERF had a more pronounced weight loss (7.1 vs. 11.5 kg; P = 0.008), as well as higher rates of esophageal (p=<0.001; OR 22.9) and tracheal stenting (p=<0.001; OR 76.8). Proximal tumor growth (p=0.004; OR 7.9), fistula formation to the trachea (p=<0.001; OR 17.2) and recurrent disease (p=0.04, OR 4.7) was associated with VFP development in EC patients with ERF. Vocal fold paralysis in ERF did not adversely affect five-year survival. CONCLUSIONS: Vocal fold paralysis is a common complication in more than half of the patients with ERF in EC. It is associated with proximal tumor growth, fistula formation to the trachea and disease recurrence, but does not influence survival.

Risk Factors for the Development of Esophagorespiratory Fistula in Esophageal Cancer.

BACKGROUND AND AIMS: The development of esophagorespiratory fistula (ERF) in esophageal cancer (EC) is a devastating complication, leading to poor survival rates and low quality of life. Goal of this study was to identify risk factors leading to fistula formation in esophageal cancer. METHODS: We identified 47 patients with malignant ERF formation in EC in a period of 10 years. Clinical characteristics were compared by univariable analysis to 47 randomly selected patients with EC, but without ERF. A case-control study was conducted for patients with squamous cell carcinoma (SCC) and ERF matching in a 1:2 fashion for primary tumor localization. RESULTS: Identifiable risk factors in EC patients were histology of SCC (P-value < 0.001), former or current smoking status (P = 0.002) and primary tumor localization in the proximal esophagus (P < 0.001). The "hot spot" for ERF formation was tumor growth 20-25cm distal to dental arch. An additional risk factor in SCC patients was age. Patients with ERF formation in SCC were younger than patients without ERF (median 63 vs. 67 years, P = 0.02). No difference in the rate of fistula formation was seen between esophagectomy and definitive chemoradiation, but the latter developed ERF earlier in the course of the disease (237 vs. 596.5 days, P = 0.01). CONCLUSION: Patients with proximal SCC of the esophagus and a smoking history, as well as young patients with SCC should be closely monitored for ERF formation.

Efficacy of palliative chemotherapy in elderly patients with colorectal cancer.

BACKGROUND: The number of old patients suffering from colorectal cancer rises. In clinical trials, old patients are underrepresented, and chemotherapy is significantly less often performed in elderly patients. We analyzed the impact of elder age for palliative chemotherapy in patients suffering from metastatic colorectal cancer, according to therapeutic drugs used, intensity of treatment performed, and therapeutic results. MATERIALS AND METHODS: We analyzed consecutive patients with metastatic colorectal cancer treated in palliative intention in our department. Assessed data included age ( 75 years), sex, comorbidity, site of primary tumor, k-ras-status, site and amount of metastasis, number and kind of chemotherapeutic agents used, number of consecutive therapy lines performed, dose intensity, toxicity, time between start and end of palliative chemotherapy, and overall survival. Prognostic variables were tested in uni- and multivariate analysis. RESULTS: Ninety-seven patients (69 < 75, 18 > 75 years) were included. Age groups were well balanced according to site of primary tumor, k-ras-mutational status, localization, and number of metastatic sites. Cardial and renal comorbidity was more frequent in elderly patients. The median number of chemotherapeutic drugs used and lines of therapy performed did not differ between age groups, except of oxaliplatin, which was significantly less often used in old patients. Median survival did not differ between age groups (23.4 vs. 23.5 months). In multivariate analysis, only left-sided primary tumor and more than 3 lines of therapy performed were prognostic positive variables. CONCLUSION: Old patients can profit from palliative chemotherapy to the same extent as younger ones.

Algorithm guided outlining of 105 pancreatic cancer liver metastases in Ultrasound.

Manual segmentation of hepatic metastases in ultrasound images acquired from patients suffering from pancreatic cancer is common practice. Semiautomatic measurements promising assistance in this process are often assessed using a small number of lesions performed by examiners who already know the algorithm. In this work, we present the application of an algorithm for the segmentation of liver metastases due to pancreatic cancer using a set of 105 different images of metastases. The algorithm and the two examiners had never assessed the images before. The examiners first performed a manual segmentation and, after five weeks, a semiautomatic segmentation using the algorithm. They were satisfied in up to 90% of the cases with the semiautomatic segmentation results. Using the algorithm was significantly faster and resulted in a median Dice similarity score of over 80%. Estimation of the inter-operator variability by using the intra class correlation coefficient was good with 0.8. In conclusion, the algorithm facilitates fast and accurate segmentation of liver metastases, comparable to the current gold standard of manual segmentation.

Interactive Outlining of Pancreatic Cancer Liver Metastases in Ultrasound Images.

Ultrasound (US) is the most commonly used liver imaging modality worldwide. Due to its low cost, it is increasingly used in the follow-up of cancer patients with metastases localized in the liver. In this contribution, we present the results of an interactive segmentation approach for liver metastases in US acquisitions. A (semi-) automatic segmentation is still very challenging because of the low image quality and the low contrast between the metastasis and the surrounding liver tissue. Thus, the state of the art in clinical practice is still manual measurement and outlining of the metastases in the US images. We tackle the problem by providing an interactive segmentation approach providing real-time feedback of the segmentation results. The approach has been evaluated with typical US acquisitions from the clinical routine, and the datasets consisted of pancreatic cancer metastases. Even for difficult cases, satisfying segmentations results could be achieved because of the interactive real-time behavior of the approach. In total, 40 clinical images have been evaluated with our method by comparing the results against manual ground truth segmentations. This evaluation yielded to an average Dice Score of 85% and an average Hausdorff Distance of 13 pixels.

Sequential endoscopic ultrasound identifies predictive variables for relapse-free follow-up after neoadjuvant chemotherapy in gastric cancer.

BACKGROUND: The accuracy of endosonographic tumor staging after neoadjuvant therapy is less reliable than in primary staging. Therefore, the value of sequential endosonographic examinations after neaodjuvant chemotherapy in gastro-esophageal cancer is discussed controversially. Previous data suggest, that endoscopic ultrasound (EUS) after neoadjuvant treatment using other variables than classic uTN-criteria may identify patients with a better prognosis. METHODS: In 67 patients with locally advanced gastric cancer treated in curative intent, we performed EUS before and after neoadjuvant chemotherapy. Endosonographic yTN-stage was compared to pathohistological yTN-stage after curative resection. The uTN-stage, yuTN-stage, maximal tumor thickness and maximal lymph node diameter as well as the shift of these variables after neoadjuvant therapy were analyzed for their usefulness to predict recurrence-free follow-up. RESULTS: Accuracy of EUS for yTN-staging after neoadjuvant therapy was poor, especially in lower tumor stages. However, three heavily correlated variables analyzed by sequential EUS could be used for the prediction of prognosis: low endosonographic tumor stage (yuT0-2) after neoadjuvant chemotherapy, a decrease of two or more steps in uT-stage and a maximal tumor thickness of <15 mm after chemotherapy were significantly associated with recurrence-free follow-up. Endosonographic T-stage before neoadjuvant therapy, as well as lymph node variables before or after chemotherapy, were of no predictive value. CONCLUSION: In spite of poor concordance between endosonographic and pathohistological TN-stage after neoadjuvant treatment, sequential EUS, performed before and after neoadjuvant therapy, possibly identify patients at risk for tumor relapse after multimodal treatment in gastric cancer. This finding should be validated in a larger patient cohort.