Role of Machine Learning in Precision Oncology: Applications in Gastrointestinal Cancers.

Gastrointestinal (GI) cancers, consisting of a wide spectrum of pathologies, have become a prominent health issue globally. Despite medical imaging playing a crucial role in the clinical workflow of cancers, standard evaluation of different imaging modalities may provide limited information. Accurate tumor detection, characterization, and monitoring remain a challenge. Progress in quantitative imaging analysis techniques resulted in "radiomics", a promising methodical tool that helps to personalize diagnosis and treatment optimization. Radiomics, a sub-field of computer vision analysis, is a bourgeoning area of interest, especially in this era of precision medicine. In the field of oncology, radiomics has been described as a tool to aid in the diagnosis, classification, and categorization of malignancies and to predict outcomes using various endpoints. In addition, machine learning is a technique for analyzing and predicting by learning from sample data, finding patterns in it, and applying it to new data. Machine learning has been increasingly applied in this field, where it is being studied in image diagnosis. This review assesses the current landscape of radiomics and methodological processes in GI cancers (including gastric, colorectal, liver, pancreatic, neuroendocrine, GI stromal, and rectal cancers). We explain in a stepwise fashion the process from data acquisition and curation to segmentation and feature extraction. Furthermore, the applications of radiomics for diagnosis, staging, assessment of tumor prognosis and treatment response according to different GI cancer types are explored. Finally, we discussed the existing challenges and limitations of radiomics in abdominal cancers and investigate future opportunities.

Percutaneous Transjejunal Biliary Access in 60 Patients with Bilioenteric Anastomoses.

PURPOSE: To determine success and complication rates of percutaneous transjejunal biliary access (PTJBA) in patients with bilioenteric anastomoses. MATERIALS AND METHODS: In a single-center, retrospective study, 169 PTJBA procedures were performed over a 13.8-y period in 60 subjects (47 male; mean age, 54.5 y). Indications for biliary interventions were cholangitis (137 cases, 45 subjects) or hyperbilirubinemia (32 cases, 18 subjects). All patients had antecolic bilioenteric anastomoses without surgical fixation to the peritoneum (liver transplantation with hepaticojejunostomy, n = 37; hepatectomy with hepaticojejunostomy, n = 8; hepaticojejunostomy only, n = 12; pancreaticoduodenectomy, n = 3). RESULTS: Initial PTJBA was successful in 140 cases (82.8%) in 35 subjects (58.3%). Twenty-one additional PTJBAs (12.4%) in 18 subjects (30.0%) were performed secondarily following a conventional transhepatic approach. Radiographic markers on the Roux-en-Y limb (P = .14, odds ratio [OR] = 2.98) or preprocedural imaging (P = .13, OR = 10.00) did not increase the odds of successful PTJBA. There were 7 major complications (4.3%) in 6 patients (10.0%) requiring hospitalization longer than 5 d, and 37 minor complications (23.0%) in 19 patients (31.7%). No procedure-related mortality was observed. Minor and major complication rates were not affected by time between bilioenteric anastomosis creation and PTJBA (P = .70, OR = 1.00; P = .62, OR = 1.00), longer dwell time of a transjejunal drain (P = .68, OR = 1.02; P = .49, OR = 0.71), or access size (P = .40, OR = 0.85; P = .23, OR = 0.59). CONCLUSIONS: PTJBA is a relatively safe technique with a high success rate in patients with bilioenteric loops that are not surgically fixed to the peritoneum.

Purposeful creation of a pneumothorax and chest tube placement to facilitate CT-guided coil localization of lung nodules before video-assisted thoracoscopic surgical wedge resection.

PURPOSE: To evaluate the feasibility and efficacy of pneumothorax creation and chest tube insertion before computed tomography (CT)-guided coil localization of small peripheral lung nodules for video-assisted thoracoscopic surgical (VATS) wedge resection. MATERIALS AND METHODS: From May 2011 to October 2013, 21 consecutive patients (seven men; mean age, 62 y; range, 42-76 y) scheduled for VATS wedge resection required CT-guided coil localization for small, likely nonpalpable peripheral lung lesions at a single institution. Outcomes were evaluated retrospectively for technical success and complications. RESULTS: There were 12 nodules and nine ground-glass opacities. Mean lesion distance from the pleural surface was 15 mm (range, 5-35 mm), and average size was 13 mm (range, 7-30 mm). A pneumothorax was successfully created in all patients with a Veress needle, and a chest tube was inserted. All target lesions were marked successfully, leaving one end of the coil within/beyond the lesion and the other end of the coil in the pleural space. The inserted chest tube was used to insufflate air to widen the pleural space during coil positioning and to aspirate any residual air before transfer of the patient to the operating room holding area. Intraparenchymal hemorrhages smaller than 7 cm in diameter developed in two patients during coil placement. All lesions were successfully resected with VATS. Histologic examinaiton revealed 13 primary adenocarcinomas, four metastases, and four benign lesions. CONCLUSIONS: Pneumothorax creation and chest tube placement before CT-guided coil localization of peripheral lung nodules for VATS wedge resection facilitates the deployment of the peripheral end of the coil in the pleural space and provides effective management of procedure-related pneumothorax until surgery.

Segmental arterial mediolysis.

Segmental arterial mediolysis (SAM) is an uncommon, nonatherosclerotic, noninflammatory, large- to medium-sized arteriopathy first described in 1976. It is characterized histologically by vacuolization and lysis of the outer arterial media leading to dissecting aneurysms and vessel rupture presenting clinically with self-limiting abdominal pain or catastrophic hemorrhages in the abdomen. Patients of all ages are affected with a greater incidence at the fifth and sixth decades. There is a slight male predominance. Imaging findings overlap with inflammatory vasculitis, collagen vascular disease, and fibromuscular dysplasia. The presence of segmental dissections involving the celiac, mesenteric, and/or renal arteries is the key distinguishing features of SAM. Inflammatory markers, genetic tests for collagen vascular disorders, and hypercoagulable studies are negative. Anti-inflammatory agents and immunosuppressants are not effective. A mortality rate of 50 % has been attributed to the acute presentation with aneurysmal rupture necessitating urgent surgical or endovascular treatments; in the absence of the acute presentation, SAM is a self-limiting disease and is treated conservatively. There are no established guidelines on medical therapy, although optimal control of blood pressure is considered the main cornerstone of medical therapy. The long-term prognosis is not known.

Intravascular ultrasound-guided inferior vena cava filter placement using a single-puncture technique in 99 patients.

PURPOSE: To assess the feasibility and safety of intravascular ultrasound (IVUS) to guide inferior vena cava (IVC) filter placement using a single venous puncture technique. MATERIALS AND METHODS: Medical records of all patients who had IVC filters placed under IVUS guidance between January 1, 2005 and December 31, 2011 were retrospectively reviewed for pertinent history, results, and complications. All filters were placed using a single venous puncture technique. RESULTS: Filters were successfully placed within the IVC in 94% of the patients with 6 malpositioned filters (4 in the iliac veins and 2 in the suprarenal IVC). Complications included groin hematoma (4%), deep venous thrombosis at the site of vascular access (2%), and filter tilt >15° along the long axis of the IVC (2%). CONCLUSIONS: The IVUS-guided IVC filter placement using a single venous puncture technique is technically feasible and safe when compared to case series using a double venous puncture technique.

Preliminary experience with Option inferior vena cava filter.

OBJECTIVES: To evaluate the safety and efficacy of Option inferior vena cava (IVC) filter during placement and short-term follow-up. METHODS: A total of 165 patients (mean age: 60-years) who received Option IVC filter from June 2009 to July 2011 were included. In all, 42 patients presented with deep vein thrombosis (DVT), 26 with pulmonary embolism (PE), and 17 with both. All outcomes were examined until April 30, 2012. RESULTS: The filters were successfully deployed in 161patients. During follow-up (mean, 9.5 ± 0.68months), 10 patients were diagnosed with post-filter PE and 13 patients with DVT. There were no instances of fatal PE. Follow-up abdominal computed tomography was available in 60 patients and demonstrated filter-related problems in 8 patients (2: penetration of filter legs, 5: asymptomatic nonocclusive thrombus, and 1: caval occlusion). There were no instances of filter migration or fracture. In total, 27 filters were successfully retrieved after a mean of 5.27 ± 0.76 months. CONCLUSION: The Option filter was effective and safe during implantation and short-term follow-up and associated with high technical success at retrieval.

Transarterial chemoembolization with Doxorubicin-eluting microspheres for inoperable hepatocellular carcinoma.

BACKGROUND: Therapy with drug-eluting microspheres was recently introduced with an aim to decrease the high postoperative morbidity associated with chemoembolization with lipiodol. The purpose of our study was to assess the safety and efficacy of chemoembolization with doxorubicin-eluting microspheres (DEB-TACE) for inoperable hepatocellular carcinoma (HCC). MATERIAL AND METHODS: In this IRB-approved retrospective study, 54 patients (44 men; median age, 61 years) with inoperable HCC were treated with DEB-TACE. HCC was diagnosed by biopsy in 43 and with a combination of α-fetoprotein (AFP) and imaging in 11. Patients with Child-Pugh A, B, C cirrhosis were 27, 25, 2, respectively. Twenty-one patients had received local therapies prior to DEB-TACE. Tumor was multifocal in 30. Eight patients had branch portal vein thrombosis. DEB-TACE was performed using 300-500µ LC Beads™ mixed with 100 mg of doxorubicin. Twenty-two patients had one DEB-TACE procedure, 23 patients had 2, 8 patients had 3, and 1 had four procedures. Response rate (RR) was assessed using AFP, RECIST, and EASL criteria on CT/MRI at 1 and 3 months. Overall median survival and survival rates at 6, 12, and 24 months were calculated. RESULTS: DEB-TACE was technically successful in all. Mean hospital stay after the procedure was 1.59 days. Thirty-day mortality was 0%. RR based on AFP was 26%. At 1 and 3 months, CR + PR were 14.8% and 35%, SD 74.1% and 50%, and PD 11.1% and 15%. Overall median survival was 445 days (95% CI 312-590). The survival rates at 6 months, 1 year, and 2 years were 77%, 59%, and 32% respectively. CONCLUSIONS: Chemoembolization with doxorubicin-eluting microspheres is safe and well tolerated in patients with inoperable HCC. Its efficacy is comparable to the historical controls. However, further prospective studies are required to confirm its efficacy.