Identifying Genetic Mutation Status in Patients with Colorectal Cancer Liver Metastases Using Radiomics-Based Machine-Learning Models.

For patients with colorectal cancer liver metastases (CRLM), the genetic mutation status is important in treatment selection and prognostication for survival outcomes. This study aims to investigate the relationship between radiomics imaging features and the genetic mutation status (KRAS mutation versus no mutation) in a large multicenter dataset of patients with CRLM and validate these findings in an external dataset. Patients with initially unresectable CRLM treated with systemic therapy of the randomized controlled CAIRO5 trial (NCT02162563) were included. All CRLM were semi-automatically segmented in pre-treatment CT scans and radiomics features were calculated from these segmentations. Additionally, data from the Netherlands Cancer Institute (NKI) were used for external validation. A total of 255 patients from the CAIRO5 trial were included. Random Forest, Gradient Boosting, Gradient Boosting + LightGBM, and Ensemble machine-learning classifiers showed AUC scores of 0.77 (95%CI 0.62-0.92), 0.77 (95%CI 0.64-0.90), 0.72 (95%CI 0.57-0.87), and 0.86 (95%CI 0.76-0.95) in the internal test set. Validation of the models on the external dataset with 129 patients resulted in AUC scores of 0.47-0.56. Machine-learning models incorporating CT imaging features could identify the genetic mutation status in patients with CRLM with a good accuracy in the internal test set. However, in the external validation set, the models performed poorly. External validation of machine-learning models is crucial for the assessment of clinical applicability and should be mandatory in all future studies in the field of radiomics.

Deep learning models for automatic tumor segmentation and total tumor volume assessment in patients with colorectal liver metastases.

BACKGROUND: We developed models for tumor segmentation to automate the assessment of total tumor volume (TTV) in patients with colorectal liver metastases (CRLM). METHODS: In this prospective cohort study, pre- and post-systemic treatment computed tomography (CT) scans of 259 patients with initially unresectable CRLM of the CAIRO5 trial (NCT02162563) were included. In total, 595 CT scans comprising 8,959 CRLM were divided into training (73%), validation (6.5%), and test sets (21%). Deep learning models were trained with ground truth segmentations of the liver and CRLM. TTV was calculated based on the CRLM segmentations. An external validation cohort was included, comprising 72 preoperative CT scans of patients with 112 resectable CRLM. Image segmentation evaluation metrics and intraclass correlation coefficient (ICC) were calculated. RESULTS: In the test set (122 CT scans), the autosegmentation models showed a global Dice similarity coefficient (DSC) of 0.96 (liver) and 0.86 (CRLM). The corresponding median per-case DSC was 0.96 (interquartile range [IQR] 0.95-0.96) and 0.80 (IQR 0.67-0.87). For tumor segmentation, the intersection-over-union, precision, and recall were 0.75, 0.89, and 0.84, respectively. An excellent agreement was observed between the reference and automatically computed TTV for the test set (ICC 0.98) and external validation cohort (ICC 0.98). In the external validation, the global DSC was 0.82 and the median per-case DSC was 0.60 (IQR 0.29-0.76) for tumor segmentation. CONCLUSIONS: Deep learning autosegmentation models were able to segment the liver and CRLM automatically and accurately in patients with initially unresectable CRLM, enabling automatic TTV assessment in such patients. RELEVANCE STATEMENT: Automatic segmentation enables the assessment of total tumor volume in patients with colorectal liver metastases, with a high potential of decreasing radiologist's workload and increasing accuracy and consistency. KEY POINTS: • Tumor response evaluation is time-consuming, manually performed, and ignores total tumor volume. • Automatic models can accurately segment tumors in patients with colorectal liver metastases. • Total tumor volume can be accurately calculated based on automatic segmentations.

Interobserver Variability in CT-based Morphologic Tumor Response Assessment of Colorectal Liver Metastases.

Purpose To evaluate interobserver variability in the morphologic tumor response assessment of colorectal liver metastases (CRLM) managed with systemic therapy and to assess the relation of morphologic response with gene mutation status, targeted therapy, and Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 measurements. Materials and Methods Participants with initially unresectable CRLM receiving different systemic therapy regimens from the randomized, controlled CAIRO5 trial (NCT02162563) were included in this prospective imaging study. Three radiologists independently assessed morphologic tumor response on baseline and first follow-up CT scans according to previously published criteria. Two additional radiologists evaluated disagreement cases. Interobserver agreement was calculated by using Fleiss κ. On the basis of the majority of individual radiologic assessments, the final morphologic tumor response was determined. Finally, the relation of morphologic tumor response and clinical prognostic parameters was assessed. Results In total, 153 participants (median age, 63 years [IQR, 56-71]; 101 men) with 306 CT scans comprising 2192 CRLM were included. Morphologic assessment performed by the three radiologists yielded 86 (56%) agreement cases and 67 (44%) disagreement cases (including four major disagreement cases). Overall interobserver agreement between the panel radiologists on morphology groups and morphologic response categories was moderate (κ = 0.53, 95% CI: 0.48, 0.58 and κ = 0.54, 95% CI: 0.47, 0.60). Optimal morphologic response was particularly observed in patients treated with bevacizumab (P = .001) and in patients with RAS/BRAF mutation (P = .04). No evidence of a relationship between RECIST 1.1 and morphologic response was found (P = .61). Conclusion Morphologic tumor response assessment following systemic therapy in participants with CRLM demonstrated considerable interobserver variability. Keywords: Tumor Response, Observer Performance, CT, Liver, Metastases, Oncology, Abdomen/Gastrointestinal Clinical trial registration no. NCT02162563 Supplemental material is available for this article. © RSNA, 2022.

Optimizing discharge after major surgery using an artificial intelligence-based decision support tool (DESIRE): An external validation study.

BACKGROUND: In the DESIRE study (Discharge aftEr Surgery usIng aRtificial intElligence), we have previously developed and validated a machine learning concept in 1,677 gastrointestinal and oncology surgery patients that can predict safe hospital discharge after the second postoperative day. Despite strong model performance (area under the receiver operating characteristics curve of 0.88) in an academic surgical population, it remains unknown whether these findings can be translated to other hospitals and surgical populations. We therefore aimed to determine the generalizability of the previously developed machine learning concept. METHODS: We externally validated the machine learning concept in gastrointestinal and oncology surgery patients admitted to 3 nonacademic hospitals in The Netherlands between January 2017 and June 2021, who remained admitted 2 days after surgery. Primary outcome was the ability to predict hospital interventions after the second postoperative day, which were defined as unplanned reoperations, radiological interventions, and/or intravenous antibiotics administration. Four forest models were locally trained and evaluated with respect to area under the receiver operating characteristics curve, sensitivity, specificity, positive predictive value, and negative predictive value. RESULTS: All models were trained on 1,693 epsiodes, of which 731 (29.9%) required a hospital intervention and demonstrated strong performance (area under the receiver operating characteristics curve only varied 4%). The best model achieved an area under the receiver operating characteristics curve of 0.83 (95% confidence interval [0.81-0.85]), sensitivity of 77.9% (0.67-0.87), specificity of 79.2% (0.72-0.85), positive predictive value of 61.6% (0.54-0.69), and negative predictive value of 89.3% (0.85-0.93). CONCLUSION: This study showed that a previously developed machine learning concept can predict safe discharge in different surgical populations and hospital settings (academic versus nonacademic) by training a model on local patient data. Given its high accuracy, integration of the machine learning concept into the clinical workflow could expedite surgical discharge and aid hospitals in addressing capacity challenges by reducing avoidable bed-days.

Developing, implementing and governing artificial intelligence in medicine: a step-by-step approach to prevent an artificial intelligence winter.

OBJECTIVE: Although the role of artificial intelligence (AI) in medicine is increasingly studied, most patients do not benefit because the majority of AI models remain in the testing and prototyping environment. The development and implementation trajectory of clinical AI models are complex and a structured overview is missing. We therefore propose a step-by-step overview to enhance clinicians' understanding and to promote quality of medical AI research. METHODS: We summarised key elements (such as current guidelines, challenges, regulatory documents and good practices) that are needed to develop and safely implement AI in medicine. CONCLUSION: This overview complements other frameworks in a way that it is accessible to stakeholders without prior AI knowledge and as such provides a step-by-step approach incorporating all the key elements and current guidelines that are essential for implementation, and can thereby help to move AI from bytes to bedside.

Imaging-based Machine-learning Models to Predict Clinical Outcomes and Identify Biomarkers in Pancreatic Cancer: A Scoping Review.

OBJECTIVE: To perform a scoping review of imaging-based machine-learning models to predict clinical outcomes and identify biomarkers in patients with PDAC. SUMMARY OF BACKGROUND DATA: Patients with PDAC could benefit from better selection for systemic and surgical therapy. Imaging-based machine-learning models may improve treatment selection. METHODS: A scoping review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses-scoping review guidelines in the PubMed and Embase databases (inception-October 2020). The review protocol was prospectively registered (open science framework registration: m4cyx). Included were studies on imaging-based machine-learning models for predicting clinical outcomes and identifying biomarkers for PDAC. The primary outcome was model performance. An area under the curve (AUC) of ≥0.75, or a P-value of ≤0.05, was considered adequate model performance. Methodological study quality was assessed using the modified radiomics quality score. RESULTS: After screening 1619 studies, 25 studies with 2305 patients fulfilled the eligibility criteria. All but 1 study was published in 2019 and 2020. Overall, 23/25 studies created models using radiomics features, 1 study quantified vascular invasion on computed tomography, and one used histopathological data. Nine models predicted clinical outcomes with AUC measures of 0.78-0.95, and C-indices of 0.65-0.76. Seventeen models identified biomarkers with AUC measures of 0.68-0.95. Adequate model performance was reported in 23/25 studies. The methodological quality of the included studies was suboptimal, with a median modified radiomics quality score score of 7/36. CONCLUSIONS: The use of imaging-based machine-learning models to predict clinical outcomes and identify biomarkers in patients with PDAC is increasingly rapidly. Although these models mostly have good performance scores, their methodological quality should be improved.

KRAS A146 Mutations Are Associated With Distinct Clinical Behavior in Patients With Colorectal Liver Metastases.

Somatic KRAS mutations occur in approximately half of the patients with metastatic colorectal cancer (mCRC). Biologic tumor characteristics differ on the basis of the KRAS mutation variant. KRAS mutations are known to influence patient prognosis and are used as predictive biomarker for treatment decisions. This study examined clinical features of patients with mCRC with a somatic mutation in KRAS G12, G13, Q61, K117, or A146. METHODS: A total of 419 patients with colorectal cancer with initially unresectable liver-limited metastases, who participated in a multicenter prospective trial, were evaluated for tumor tissue KRAS mutation status. For the subgroup of patients who carried a KRAS mutation and were treated with bevacizumab and doublet or triplet chemotherapy (N = 156), pretreatment circulating tumor DNA levels were analyzed, and total tumor volume (TTV) was quantified on the pretreatment computed tomography images. RESULTS: Most patients carried a KRAS G12 mutation (N = 112), followed by mutations in G13 (N = 15), A146 (N = 12), Q61 (N = 9), and K117 (N = 5). High plasma circulating tumor DNA levels were observed for patients carrying a KRAS A146 mutation versus those with a KRAS G12 mutation, with median mutant allele frequencies of 48% versus 19%, respectively. Radiologic TTV revealed this difference to be associated with a higher tumor load in patients harboring a KRAS A146 mutation (median TTV 672 cm3 [A146] v 74 cm3 [G12], P = .036). Moreover, KRAS A146 mutation carriers showed inferior overall survival compared with patients with mutations in KRAS G12 (median 10.7 v 26.4 months; hazard ratio = 2.5; P = .003). CONCLUSION: Patients with mCRC with a KRAS A146 mutation represent a distinct molecular subgroup of patients with higher tumor burden and worse clinical outcomes, who might benefit from more intensive treatments. These results highlight the importance of testing colorectal cancer for all KRAS mutations in routine clinical care.

Artificial Intelligence-Based Segmentation of Residual Tumor in Histopathology of Pancreatic Cancer after Neoadjuvant Treatment.

BACKGROUND: Histologic examination of resected pancreatic cancer after neoadjuvant therapy (NAT) is used to assess the effect of NAT and may guide the choice for adjuvant treatment. However, evaluating residual tumor burden in pancreatic cancer is challenging given tumor response heterogeneity and challenging histomorphology. Artificial intelligence techniques may offer a more reproducible approach. METHODS: From 64 patients, one H&E-stained slide of resected pancreatic cancer after NAT was digitized. Three separate classes were manually outlined in each slide (i.e., tumor, normal ducts, and remaining epithelium). Corresponding segmentation masks and patches were generated and distributed over training, validation, and test sets. Modified U-nets with varying encoders were trained, and F1 scores were obtained to express segmentation accuracy. RESULTS: The highest mean segmentation accuracy was obtained using modified U-nets with a DenseNet161 encoder. Tumor tissue was segmented with a high mean F1 score of 0.86, while the overall multiclass average F1 score was 0.82. CONCLUSIONS: This study shows that artificial intelligence-based assessment of residual tumor burden is feasible given the promising obtained F1 scores for tumor segmentation. This model could be developed into a tool for the objective evaluation of the response to NAT and may potentially guide the choice for adjuvant treatment.

Generating insights in uncharted territories: real-time learning from data in critically ill patients-an implementer report.

Introduction In the current situation, clinical patient data are often siloed in multiple hospital information systems. Especially in the intensive care unit (ICU), large volumes of clinical data are routinely collected through continuous patient monitoring. Although these data often contain useful information for clinical decision making, they are not frequently used to improve quality of care. During, but also after, pressing times, data-driven methods can be used to mine treatment patterns from clinical data to determine the best treatment options from a hospitals own clinical data.Methods In this implementer report, we describe how we implemented a data infrastructure that enabled us to learn in real time from consecutive COVID-19 ICU admissions. In addition, we explain our step-by-step multidisciplinary approach to establish such a data infrastructure.Conclusion By sharing our steps and approach, we aim to inspire others, in and outside ICU walls, to make more efficient use of data at hand, now and in the future.

Postoperative circulating tumour DNA is associated with pathologic response and recurrence-free survival after resection of colorectal cancer liver metastases.

BACKGROUND: Recurrence rates after resection of colorectal cancer liver metastases (CRLM) are high and correlate with worse survival. Postoperative circulating tumour DNA (ctDNA) is a promising prognostic biomarker. Focusing on patients with resected CRLM, this study aimed to evaluate the association between the detection of postoperative ctDNA, pathologic response and recurrence-free survival (RFS). METHODS: Twenty-three patients were selected from an ongoing phase-3 trial who underwent resection of RAS-mutant CRLM after induction systemic treatment. CtDNA analysis was performed by droplet digital PCR using blood samples collected at baseline, before and after resection. Pathologic response of CRLM was determined via the Tumour Regression Grading system. FINDINGS: With a median follow-up of 19.6 months, the median RFS for patients with detectable (N = 6, [26%]) and undetectable (N = 17, [74%]) postoperative ctDNA was 4.8 versus 12.1 months, respectively. Among 21 patients with available tumour tissue, pathologic response in patients with detectable compared to undetectable postoperative ctDNA was found in one of six (17%) and 15 of 15 (100%) patients, respectively (p < 0.001). In univariable Cox regression analyses both postoperative detectable ctDNA (HR = 3.3, 95%CI = 1.1-9.6, p = 0.03) and pathologic non-response (HR = 4.6, 95%CI = 1.4-15, p = 0.01) were associated with poorer RFS and were strongly correlated (r = 0.88, p < 0.001). After adjusting for clinical characteristics in pairwise multivariable analyses, postoperative ctDNA status remained associated with RFS. INTERPRETATION: The detection of postoperative ctDNA after secondary resection of CRLM is a promising prognostic factor for RFS and appeared to be highly correlated with pathologic response. FUNDING: None.

Predicting need for hospital-specific interventional care after surgery using electronic health record data.

BACKGROUND: A significant proportion of surgical inpatients is often admitted longer than necessary. Early identification of patients who do not need care that is strictly provided within hospitals would allow timely discharge of patients to a postoperative nursing home for further recovery. We aimed to develop a model to predict whether a patient needs hospital-specific interventional care beyond the second postoperative day. METHODS: This study included all adult patients discharged from surgical care in the surgical oncology department from June 2017 to February 2020. The primary outcome was to predict whether a patient still needs hospital-specific interventional care beyond the second postoperative day. Hospital-specific care was defined as unplanned reoperations, radiological interventions, and intravenous antibiotics administration. Different analytical methods were compared with respect to the area under the receiver-operating characteristics curve, sensitivity, specificity, positive predictive value, and negative predictive value. RESULTS: Each model was trained on 1,174 episodes. In total, 847 (50.5%) patients required an intervention during postoperative admission. A random forest model performed best with an area under the receiver-operating characteristics curve of 0.88 (95% confidence interval 0.83-0.93), sensitivity of 79.1% (95% confidence interval 0.67-0.92), specificity of 80.0% (0.73-0.87), positive predictive value of 57.6% (0.45-0.70) and negative predictive value of 91.7% (0.87-0.97). CONCLUSION: This proof-of-concept study found that a random forest model could successfully predict whether a patient could be safely discharged to a nursing home and does not need hospital care anymore. Such a model could aid hospitals in addressing capacity challenges and improve patient flow, allowing for timely surgical care.

Moving from bytes to bedside: a systematic review on the use of artificial intelligence in the intensive care unit.

PURPOSE: Due to the increasing demand for intensive care unit (ICU) treatment, and to improve quality and efficiency of care, there is a need for adequate and efficient clinical decision-making. The advancement of artificial intelligence (AI) technologies has resulted in the development of prediction models, which might aid clinical decision-making. This systematic review seeks to give a contemporary overview of the current maturity of AI in the ICU, the research methods behind these studies, and the risk of bias in these studies. METHODS: A systematic search was conducted in Embase, Medline, Web of Science Core Collection and Cochrane Central Register of Controlled Trials databases to identify eligible studies. Studies using AI to analyze ICU data were considered eligible. Specifically, the study design, study aim, dataset size, level of validation, level of readiness, and the outcomes of clinical trials were extracted. Risk of bias in individual studies was evaluated by the Prediction model Risk Of Bias ASsessment Tool (PROBAST). RESULTS: Out of 6455 studies identified through literature search, 494 were included. The most common study design was retrospective [476 studies (96.4% of all studies)] followed by prospective observational [8 (1.6%)] and clinical [10 (2%)] trials. 378 (80.9%) retrospective studies were classified as high risk of bias. No studies were identified that reported on the outcome evaluation of an AI model integrated in routine clinical practice. CONCLUSION: The vast majority of developed ICU-AI models remain within the testing and prototyping environment; only a handful were actually evaluated in clinical practice. A uniform and structured approach can support the development, safe delivery, and implementation of AI to determine clinical benefit in the ICU.

Advanced analytics and artificial intelligence in gastrointestinal cancer: a systematic review of radiomics predicting response to treatment.

PURPOSE: Advanced medical image analytics is increasingly used to predict clinical outcome in patients diagnosed with gastrointestinal tumors. This review provides an overview on the value of radiomics in predicting response to treatment in patients with gastrointestinal tumors. METHODS: A systematic review was conducted, according to PRISMA guidelines. The protocol was prospectively registered (PROSPERO: CRD42019128408). PubMed, Embase, and Cochrane databases were searched. Original studies reporting on the value of radiomics in predicting response to treatment in patients with a gastrointestinal tumor were included. A narrative synthesis of results was conducted. Results were stratified by tumor type. Quality assessment of included studies was performed, according to the radiomics quality score. RESULTS: The comprehensive literature search identified 1360 unique studies, of which 60 articles were included for analysis. In 37 studies, radiomics models and individual radiomic features showed good predictive performance for response to treatment (area under the curve or accuracy > 0.75). Various strategies to construct predictive models were used. Internal validation of predictive models was often performed, while the majority of studies lacked external validation. None of the studies reported predictive models implemented in clinical practice. CONCLUSION: Radiomics is increasingly used to predict response to treatment in patients suffering from gastrointestinal cancer. This review demonstrates its great potential to help predict response to treatment and improve patient selection and early adjustment of treatment strategy in a non-invasive manner.

Short-Term Outcomes of Secondary Liver Surgery for Initially Unresectable Colorectal Liver Metastases Following Modern Induction Systemic Therapy in the Dutch CAIRO5 Trial.

Objective: To present short-term outcomes of liver surgery in patients with initially unresectable colorectal liver metastases (CRLM) downsized by chemotherapy plus targeted agents. Background: The increase of complex hepatic resections of CRLM, technical innovations pushing boundaries of respectability, and use of intensified induction systemic regimens warrant for safety data in a homogeneous multicenter prospective cohort. Methods: Patients with initially unresectable CRLM, who underwent complete resection after induction systemic regimens with doublet or triplet chemotherapy, both plus targeted therapy, were selected from the ongoing phase III CAIRO5 study (NCT02162563). Short-term outcomes and risk factors for severe postoperative morbidity (Clavien Dindo grade ≥ 3) were analyzed using logistic regression analysis. Results: A total of 173 patients underwent resection of CRLM after induction systemic therapy. The median number of metastases was 9 and 161 (93%) patients had bilobar disease. Thirty-six (20.8%) 2-stage resections and 88 (51%) major resections (>3 liver segments) were performed. Severe postoperative morbidity and 90-day mortality was 15.6% and 2.9%, respectively. After multivariable analysis, blood transfusion (odds ratio [OR] 2.9 [95% confidence interval (CI) 1.1-6.4], P = 0.03), major resection (OR 2.9 [95% CI 1.1-7.5], P = 0.03), and triplet chemotherapy (OR 2.6 [95% CI 1.1-7.5], P = 0.03) were independently correlated with severe postoperative complications. No association was found between number of cycles of systemic therapy and severe complications (r = -0.038, P = 0.31). Conclusion: In patients with initially unresectable CRLM undergoing modern induction systemic therapy and extensive liver surgery, severe postoperative morbidity and 90-day mortality were 15.6% and 2.7%, respectively. Triplet chemotherapy, blood transfusion, and major resections were associated with severe postoperative morbidity.

The Prognostic Value of Total Tumor Volume Response Compared With RECIST1.1 in Patients With Initially Unresectable Colorectal Liver Metastases Undergoing Systemic Treatment.

Objectives: Compare total tumor volume (TTV) response after systemic treatment to Response Evaluation Criteria in Solid Tumors (RECIST1.1) and assess the prognostic value of TTV change and RECIST1.1 for recurrence-free survival (RFS) in patients with colorectal liver-only metastases (CRLM). Background: RECIST1.1 provides unidimensional criteria to evaluate tumor response to systemic therapy. Those criteria are accepted worldwide but are limited by interobserver variability and ignore potentially valuable information about TTV. Methods: Patients with initially unresectable CRLM receiving systemic treatment from the randomized, controlled CAIRO5 trial (NCT02162563) were included. TTV response was assessed using software specifically developed together with SAS analytics. Baseline and follow-up computed tomography (CT) scans were used to calculate RECIST1.1 and TTV response to systemic therapy. Different thresholds (10%, 20%, 40%) were used to define response of TTV as no standard currently exists. RFS was assessed in a subgroup of patients with secondarily resectable CRLM after induction treatment. Results: A total of 420 CT scans comprising 7820 CRLM in 210 patients were evaluated. In 30% to 50% (depending on chosen TTV threshold) of patients, discordance was observed between RECIST1.1 and TTV change. A TTV decrease of >40% was observed in 47 (22%) patients who had stable disease according to RECIST1.1. In 118 patients with secondarily resectable CRLM, RFS was shorter for patients with less than 10% TTV decrease compared with patients with more than 10% TTV decrease (P = 0.015), while RECIST1.1 was not prognostic (P = 0.821). Conclusions: TTV response assessment shows prognostic potential in the evaluation of systemic therapy response in patients with CRLM.

Predicting thromboembolic complications in COVID-19 ICU patients using machine learning.

BACKGROUND: The coronavirus disease 2019 (COVID-19) pandemic is a challenge for intensive care units (ICU) in part due to the failure to identify risks for patients early and the inability to render an accurate prognosis. Previous reports suggest a strong association between hypercoagulability and poor outcome. Factors related to hemostasis may, therefore, serve as tools to improve the management of COVID-19 patients. AIM: The purpose of this report is to develop a model to determine whether it is possible to early identify COVID-19 patients at risk for thromboembolic complications (TCs). METHODS: We analyzed electronic health record data of 108 consecutive COVID-19 patients admitted to the adult ICU of the Erasmus University Medical Center between February 27 and May 20, 2020. By training a decision tree classifier on 66% of the available data, a model for the prediction of TCs was developed. RESULTS: The median (interquartile range) age was 62 (53-70) years and 73% were male. Forty-three patients (40%) developed a TC during their ICU stay. Mortality was higher for patients in the TCs group compared to the control group (26% vs. 8%, P=0.03). Lactate dehydrogenase, standardized bicarbonate, albumin, and leukocytes were identified by the Decision Tree classifier as the most powerful predictors for TCs 2 days before the onset of the TC, with a sensitivity of 73% and a positive likelihood ratio of 2.7 on the test dataset. CONCLUSIONS: Clinically relevant TCs frequently occur in critically ill COVID-19 patients. These can successfully be predicted using a decision tree model. Although this model could be of special importance to aid clinical decision making, its generalizability and clinical impact should be determined in a larger population. RELEVANCE FOR PATIENTS: Recently, severe TCs were observed in COVID-19 patients with progressive respiratory failure warranting ICU treatment. Timely identification of patients at risk of developing TCs is critical inasmuch as it would enable clinicians to initiate potentially salvaging therapeutic anticoagulation.

From registration to publication: A study on Dutch academic randomized controlled trials.

INTRODUCTION: Registration of clinical trials has been initiated in order to assess adherence of the reported results to the original trial protocol. This study aimed to investigate the publication rates, timely dissemination of results, and the prevalence of consistency in hypothesis, sample size, and primary endpoint of Dutch investigator-initiated randomized controlled clinical trials (RCTs). METHODS: All Dutch investigator-initiated RCTs with a completion date between December 31, 2010, and January 1, 2012, and registered in the Trial Register of The Netherlands database were included. PubMed was searched for the publication of these RCT results until September 2016, and the time to the publication date was calculated. Consistency in hypothesis, sample size, and primary endpoint compared with the registry data were assessed. RESULTS: The search resulted in a total of 168 Dutch investigator-initiated RCTs. In September 2016, the results of 129 (77%) trials had been published, of which 50 (39%) within 2 years after completion of accrual. Consistency in hypothesis with the original protocol was observed in 108 (84%) RCTs; in 71 trials (55%), the planned sample size was reached; and 103 trials (80%) presented the original primary endpoint. Consistency in all three parameters was observed in 50 studies (39%). CONCLUSION: This study shows that approximately one out of four Dutch investigator-initiated RCTs remains unpublished 5 years after initiation. The observed low overall consistency with the initial study outline is a matter of concern and warrants improvements in trial design and assessment of trial feasibility.

Outcomes of Resectability Assessment of the Dutch Colorectal Cancer Group Liver Metastases Expert Panel.

BACKGROUND: Decision making on optimal treatment strategy in patients with initially unresectable colorectal cancer liver metastases (CRLM) remains complex because uniform criteria for (un)resectability are lacking. This study reports on the feasibility and short-term outcomes of The Dutch Colorectal Cancer Group Liver Expert Panel. STUDY DESIGN: The Expert Panel consists of 13 hepatobiliary surgeons and 4 radiologists. Resectability assessment is performed independently by 3 randomly assigned surgeons, and CRLM are scored as resectable, potentially resectable, or permanently unresectable. In absence of consensus, 2 additional surgeons are invited for a majority consensus. Patients with potentially resectable or unresectable CRLM at baseline are evaluated every 2 months of systemic therapy. Once CRLM are considered resectable, a treatment strategy is proposed. RESULTS: Overall, 398 panel evaluations in 183 patients were analyzed. The median time to panel conclusion was 7 days (interquartile range [IQR] 5-11 days). Intersurgeon disagreement was observed in 205 (52%) evaluations, with major disagreement (resectable vs permanently unresectable) in 42 (11%) evaluations. After systemic treatment, 106 patients were considered to have resectable CRLM, 84 of whom (79%) underwent a curative procedure. R0 resection (n = 41), R0 resection in combination with ablative treatment (n = 26), or ablative treatment only (n = 4) was achieved in 67 of 84 (80%) patients. CONCLUSIONS: This study analyzed prospective resectability evaluation of patients with CRLM by a panel of radiologists and liver surgeons. The high rate of disagreement among experienced liver surgeons reflects the complexity in defining treatment strategies for CRLM and supports the use of a panel rather than a single-surgeon decision.

Avoiding postoperative mortality after ALPPS-development of a tumor-specific risk score for colorectal liver metastases.

BACKGROUND: ALPPS is a two-stage hepatectomy that induces more rapid liver growth compared to conventional strategies. This report aims to establish a risk-score to avoid adverse outcomes of ALPPS only for patients with colorectal liver metastases (CRLM) as primary indication for ALPPS. METHODS: All patients with CRLM included in the ALPPS registry were included. Risk score analysis was performed for 90-day mortality after ALPPS, defined as death within 90 days after either stage. Two risk scores were generated i.e. one for application before stage-1, and one for application before stage-2. Logistic regression analysis was performed to establish the risk-score. RESULTS: In total, 486 patients were included, of which 35 (7%) died 90 days after stage-1 or 2. In the stage-1 risk score, age ≥67 years (OR 3.7), FLR/BW ratio <0.40 (OR 2.9) and total center-volume (OR 2.4) were included. For the stage-2 score age ≥67 years (OR 3.7), FLR/BW ratio <0.40 (OR 2.8), bilirubin 5 days after stage-1 >50 µmol/L (OR 2.4), and stage-1 morbidity grade IIIA or higher (OR 6.3) were included. CONCLUSIONS: The CRLM risk-score to predict mortality after ALPPS demonstrates that older patients with small remnant livers in inexperienced centers, especially after experiencing morbidity after stage-1 have adverse outcomes. The risk score may be used to restrict ALPPS to low-risk patient populations.