Artificial neural networks versus LASSO regression for the prediction of long-term survival after surgery for invasive IPMN of the pancreas.

Prediction of long-term survival in patients with invasive intraductal papillary mucinous neoplasm (IPMN) of the pancreas may aid in patient assessment, risk stratification and personalization of treatment. This study aimed to investigate the predictive ability of artificial neural networks (ANN) and LASSO regression in terms of 5-year disease-specific survival. ANN work in a non-linear fashion, having a potential advantage in analysis of variables with complex correlations compared to regression models. LASSO is a type of regression analysis facilitating variable selection and regularization. A total of 440 patients undergoing surgical treatment for invasive IPMN of the pancreas registered in the Surveillance, Epidemiology and End Results (SEER) database between 2004 and 2016 were analyzed. The dataset was prior to analysis randomly split into a modelling and test set (7:3). The accuracy, precision and F1 score for predicting mortality were 0.82, 0.83 and 0.89, respectively for ANN with variable selection compared to 0.79, 0.85 and 0.87, respectively for the LASSO-model. ANN using all variables showed similar accuracy, precision and F1 score of 0.81, 0.85 and 0.88, respectively compared to a logistic regression analysis. McNemar´s test showed no statistical difference between the models. The models showed high and similar performance with regard to accuracy and precision for predicting 5-year survival status.

Intraductal papillary mucinous carcinoma versus pancreatic ductal adenocarcinoma: A systematic review and meta-analysis.

BACKGROUND: Previous studies have indicated that there may be a difference in tumor biology between intraductal papillary mucinous carcinoma (IPMC) and pancreatic ductal adenocarcinoma (PDAC). However, the data are still controversial. The aim of this systematic review and meta-analysis was to summarize and compare the outcome of IPMC and PDAC after surgical resection. METHODS: Studies comparing IPMC and PDAC were identified using Medline and Embase search engines. Primary outcomes of interest were survival and recurrence. Secondary outcomes were clinicopathological characteristics. Meta-analysis of data was conducted using a random-effects model. RESULTS: A total of 14 studies were included. Pooled analysis revealed an improved 5-year overall survival (OS) for IPMC compared to PDAC (OR 0.23, 95% CI 0.09-0.56). Both colloid and tubular IPMC showed improved 5-year OS compared to PDAC (OR 0.12, 95% CI 0.05-0.25 and OR 0.38, 95% CI 0.26-0.54, respectively). Median survival time ranged from 21 to 58 months in the IPMC group compared to 12-23 months in the PDAC group. No meta-analysis could be performed on recurrence or on time-to-event data. Descriptive data showed no survival difference for higher TNM stages. IPMC was more often found at a TNM-stage of 1 (OR 4.40, 95% CI 2.71-7.15) and had lower rates of lymph node spread (OR 0.43, 95% CI 0.32-0.57). CONCLUSION: Available data suggest that IPMC has a more indolent course with a better 5-year OS compared to PDAC. The histopathological features are less aggressive in IPMC. The reason may be earlier detection. However, for IPMC with higher TNM stages the survival seems to be similar to that of PDAC.

Adjuvant therapy in invasive intraductal papillary mucinous neoplasm (IPMN) of the pancreas: a systematic review.

Intraductal papillary mucinous neoplasm (IPMN) of the pancreas is a cystic tumor with a disease spectrum ranging from low-grade dysplasia to invasive carcinoma. The evidence for adjuvant treatment in invasive IPMN is limited and mostly derived from studies in conventional pancreatic ductal adenocarcinoma (PDAC). We performed a systematic review focusing on all clinical studies concerning the efficacy of adjuvant therapy in patients with invasive IPMN. We identified 8 retrospective cohort studies, using either adjuvant chemotherapy alone (n=1), adjuvant radiotherapy alone (n=1) or adjuvant chemotherapy in combination with radiation (n=6). Adjuvant therapy was associated with a survival benefit in 7 out of the 8 studies. Specific survival benefit was noted for patients with node-positive disease, higher TNM stage, positive resection margins, poor differentiation and tubular subtype. We conclude that adjuvant therapy may be beneficial in invasive IPMN, but current data suggest that it should be given selectively based on individual tumor characteristics. Further prospective, randomized studies are warranted.

High-density and targeted glycoproteomic profiling of serum proteins in pancreatic cancer and intraductal papillary mucinous neoplasm.

OBJECTIVE: Glycoproteomics is an emerging subfield of proteomics. Tumor-specific variations in protein glycosylation might be potential targets for the development of new cancer diagnostics. Here, we performed high-throughput screening and targeted verification of glycome alterations in serum samples from patients with pancreatic cancer and the precancerous lesion intraductal papillary mucinous neoplasm (IPMN). MATERIAL AND METHODS: The glycosylation profile of 1000 proteins was mapped in a discovery cohort comprising serum samples from 16 individuals, including 8 patients with pancreatic cancer and 8 healthy controls. The top 10 glycoprotein biomarker candidates with the highest signal intensity difference in glycosylation levels were evaluated in a cohort consisting of 109 serum samples, including 49 patients with resectable pancreatic cancer, 13 patients with resectable noninvasive IPMN and 47 healthy controls, using a targeted assay. RESULTS: Multivariable analysis defined sets of panels comprising CA19-9 and distinctively glycosylated proteins for discrimination between pancreatic cancer, IPMN and healthy controls. A panel including CA 19-9, IL.17E, B7.1 and DR6 gave an AUC of 0.988 at 100% sensitivity at 90% specificity for the discrimination of stage 1 pancreatic cancer and healthy controls. B7.1 was found to be a valuable biomarker for differentiating between IPMN and healthy controls, with better performance alone than CA 19-9. CONCLUSIONS: Measurement of protein glycosylation profiles in serum may aid in the early detection of pancreatic cancer and precursor lesions.

Intraductal papillary mucinous neoplasms of the pancreas - a cost-effectiveness analysis of management strategies for the branch-duct subtype.

BACKGROUND: Branch-duct intraductal papillary mucinous neoplasm (BD-IPMN) presents a clinical conundrum. Rigorous long-term surveillance or surgical resection is recommended. The economic consequences of the management have not been fully investigated. METHODS: A Markov decision model compared 4 strategies for low-risk BD-IPMN: I = upfront total pancreatectomy, II = upfront partial pancreatectomy, III = initial surveillance, IV = watchful waiting. Surveillance was based on the Swedish Guidelines for Pancreatic Cancer. Probabilities and costs were obtained from the participating unit and from the scientific literature. The incremental cost-effectiveness ratios (ICERs) were calculated and sensitivity analyses were performed by varying relevant parameters. Survival was reported in quality-adjusted life-years (QALYs). RESULTS: Strategy III was the most cost-effective strategy with an ICER of €31 682 compared to strategy IV. Strategy I was the most expensive but yielded the best QALY (9.32). Total number of years, annual risk of pancreatic cancer and annual risk of a low-risk BD-IPMN turning into a high-risk lesion had the greatest impact in the model. CONCLUSIONS: Initial surveillance seems to be the most cost-effective strategy in the management of low-risk asymptomatic BD-IPMN. However, the possibility of personalized approaches remains to be investigated.

Impact of body constitution on complications following pancreaticoduodenectomy: A retrospective cohort study.

BACKGROUND: Overweight, defined by body mass index (BMI), is correlated to complications following pancreaticoduodenectomy (PD). The aim of this study was to evaluate the impact of body constitution, measured with different anthropometric measures, and diabetes on complications following PD. MATERIALS AND METHODS: Patients who underwent PD between 2000 and 2015 at Skåne University Hospital were retrospectively included. Body mass index (BMI), body surface area (BSA) and body fat percentage (BF%) were calculated. Overweight and obesity were defined by BMI according to the WHO classification (overweight ≥25 and obesity ≥30). Values equal to or above the median value were considered as large by BSA (≥1.87) and overweight by BF% (≥29.6% (male) and ≥38.9% (female)). Main endpoints were events of postoperative pancreatic fistula (POPF), post pancreatectomy hemorrhage (PPH), delayed gastric emptying (DGE) and complications classified according to the Clavien-Dindo classification. Multivariable analysis was performed using logistic regression and a subgroup analysis on diabetic patients was performed. RESULTS: In total 328 patients were included. The incidence of POPF grades B and C was increased among overweight and large patients defined by BMI (OR 4.16; p = 0.001), BSA (OR 2.88; p = 0.018) and BF% (OR 3.94; p = 0.001). However, the risk was not increased among diabetic patients with BMI≥25 and BMI≥30. DGE and complications classified as Clavien grade ≥3 were more common in patients defined as overweight by both BMI (OR 1.72; p = 0.024 and OR 2.63; p = 0.003, respectively) and BF% (OR 2.13; p = 0.001 and OR 2.31; p = 0.009, respectively). PPH was not more frequent in overweight or large patients. CONCLUSION: Body constitution has an impact on the risk of severe complications following PD. BMI, BSA and BF% can all be used to identify risk groups. The risk of developing POPF grades B and C was significantly increased in overweight and large patients, but not in patients with coexisting diabetes.

Does next-generation sequencing of cyst fluid improve management of pancreatic cystic neoplasms?

Pancreatic cystic lesions represent a heterogeneous group of diseases ranging from benign to malignant lesions. They are increasingly being detected due to the widespread use of cross-sectional imaging. Their management is a challenge because it is often not possible to reliably discriminate between malignant and nonmalignant lesions using current imaging technology. Next-generation sequencing (NGS) has the ability of both whole genome and targeted sequencing at a low cost and from a limited amount of DNA. NGS of cyst fluid aspired by endoscopic ultrasonography-guided fine-needle aspiration provides a valuable tool in biomarker research and may in the future help improve diagnosis and management of pancreatic cystic lesions.

Intraductal papillary mucinous neoplasm of the pancreas - epidemiology, risk factors, diagnosis, and management.

Intraductal papillary mucinous neoplasm (IPMN) is one of the most common cystic neoplasms of the pancreas. It is a heterogeneous disease and can be divided into ductal types and morphological subtypes. The incidence of IPMN is increasing, likely due to the widespread use of cross-sectional imaging and a growing elderly population. IPMN poses an increasing demand on the health care system. Current guidelines provide indications for surgery and recommendations for surveillance, but management of IPMN is still challenging in routine clinical practice. In this article, we review current knowledge about IPMN and provide future directions for improving diagnosis and management.

Safety of pancreatic resection in the elderly: a retrospective analysis of 556 patients.

BACKGROUND: The safety of pancreatic resection for elderly patients is still controversial. We examined the postoperative morbidity and mortality in patients aged 75 years or more undergoing pancreatic resection. METHODS: Patients undergoing pancreatic resection were studied retrospectively and the outcomes were compared between patients aged <75 and ≥75 years. RESULTS: Of the 556 patients enrolled, 78 (14%) were ≥75 years old. Elderly patients had significantly more co-morbidities, especially cardiovascular pathology (P=0.005). Also, elderly patients had significantly lower body mass index prior to surgery (P=0.005). There were no significant differences in terms of surgical procedures and tumor types between age groups. The incidence of postoperative pancreatic fistula grade A was significantly lower in the elderly group (P=0.022), but no significant differences were noted in the overall morbidity or the incidence of postpancreatectomy hemorrhage, delayed gastric emptying, bile leakage, cardiac complications, pulmonary complications or septic complications. The 30-day mortality rate was similar between groups (0.8% vs. 1.3%; P=0.532). CONCLUSION: Pancreatic resection is a safe option for selected elderly patients. Our study confirms that age alone should not preclude potentially curative surgical therapy.

The role of quantitative mass spectrometry in the discovery of pancreatic cancer biomarkers for translational science.

In the post-genomic era, it has become evident that genetic changes alone are not sufficient to understand most disease processes including pancreatic cancer. Genome sequencing has revealed a complex set of genetic alterations in pancreatic cancer such as point mutations, chromosomal losses, gene amplifications and telomere shortening that drive cancerous growth through specific signaling pathways. Proteome-based approaches are important complements to genomic data and provide crucial information of the target driver molecules and their post-translational modifications. By applying quantitative mass spectrometry, this is an alternative way to identify biomarkers for early diagnosis and personalized medicine. We review the current quantitative mass spectrometric technologies and analyses that have been developed and applied in the last decade in the context of pancreatic cancer. Examples of candidate biomarkers that have been identified from these pancreas studies include among others, asporin, CD9, CXC chemokine ligand 7, fibronectin 1, galectin-1, gelsolin, intercellular adhesion molecule 1, insulin-like growth factor binding protein 2, metalloproteinase inhibitor 1, stromal cell derived factor 4, and transforming growth factor beta-induced protein. Many of these proteins are involved in various steps in pancreatic tumor progression including cell proliferation, adhesion, migration, invasion, metastasis, immune response and angiogenesis. These new protein candidates may provide essential information for the development of protein diagnostics and targeted therapies. We further argue that new strategies must be advanced and established for the integration of proteomic, transcriptomic and genomic data, in order to enhance biomarker translation. Large scale studies with meta data processing will pave the way for novel and unexpected correlations within pancreatic cancer, that will benefit the patient, with targeted treatment.