30-day mortality after elective colorectal surgery can reasonably be predicted

Outcome variation and the role of caseload in certified colorectal cancer centers - a retrospective cohort analysis of 90 000 cases

BACKGROUND

Studies have shown that surgical treatment of colorectal carcinomas in certified centers leads to improved outcomes. However, there were considerable fluctuations in outcome parameters. It has not yet been examined whether this variability is due to continuous differences between hospitals or variability within a hospital over time.

MATERIALS AND METHODS

In this retrospective observational cohort study, administrative quality assurance data of 153 German-certified colorectal cancer centers between 2010 and 2019 were analyzed. Six outcome quality indicators (QIs) were studied: 30-day postoperative mortality (POM) rate, surgical site infection (SSI) rate, anastomotic insufficiency (AI) rate, and revision surgery (RS) rate. AI and RS were also analyzed for colon (C) and rectal cancer operations (R). Variability was analyzed by funnel plots with 95% and 99% control limits and modified Cleveland dot plots.

RESULTS

In the 153 centers, 90 082 patients with colon cancer and 47 623 patients with rectal cancer were treated. Average QI scores were 2.7% POM, 6.2% SSI, 4.8% AI-C, 8.5% AI-R, 9.1% RS-C, and 9.8% RS-R. The funnel plots revealed that for every QI, about 10.1% of hospitals lay above the upper 99% and about 8.7% below the lower 99% control limit. In POM, SSI, and AI-R, a significant negative correlation with the average annual caseload was observed.

CONCLUSION

The analysis showed high variability in outcome quality between and within the certified colorectal cancer centers. Only a small number of hospitals had a high performance on all six QIs, suggesting that significant quality variation exists even within the group of certified centers.

Management of non-curative endoscopic resection of T1 colon cancer

Endoscopic resection techniques enable en-bloc resection of T1 colon cancers. A complete removal of T1 colon cancer can be considered curative when histologic examination of the specimens shows none of the high-risk factors for lymph nodes metastases. Criteria predicting lymph nodes metastases include deep submucosal invasion, poor differentiation, lymphovascular invasion, and high-grade tumor budding. In these cases, complete (R0), local endoscopic resection is considered sufficient as negligible risk of lymph nodes metastases does not outweigh morbidity and mortality associated with surgical resection. Challenges arise when endoscopic resection is incomplete (RX/R1) or high-risk histological features are present. The risk of lymph node metastasis in T1 CRC ranges from 1% to 36.4%, depending on histologic risk factors. Presence of any risk factor labels the patient "high risk," warranting oncologic surgery with mesocolic lymphadenectomy. However, even if 70%-80% of T1-CRC patients are classified as high-risk, more than 90% are without lymph node involvement after oncological surgery. Surgical overtreatment in T1 CRC is a challenge, requiring a balance between oncologic safety and minimizing morbidity/mortality. This narrative review explores the landscape of managing non-curative T1 colon cancer, focusing on the choice between advanced endoscopic resection techniques and surgical interventions. We discuss surveillance strategies and shared decision-making, emphasizing the importance of a multidisciplinary approach.

Machine learning-based prediction of postoperative mortality in emergency colorectal surgery: A retrospective, multicenter cohort study using Tokushukai medical database

Background

Although prognostic factors associated with mortality in patients with emergency colorectal surgery have been identified, an accurate mortality risk assessment is still necessary to determine the range of therapeutic resources in accordance with the severity of patients. We established machine-learning models to predict in-hospital mortality for patients who had emergency colorectal surgery using clinical data at admission and attempted to identify prognostic factors associated with in-hospital mortality.

Methods

This retrospective cohort study included adult patients undergoing emergency colorectal surgery in 42 hospitals between 2012 and 2020. We employed logistic regression and three supervised machine-learning models: random forests, gradient-boosting decision trees (GBDT), and multilayer perceptron (MLP). The area under the receiver operating characteristics curve (AUROC) was calculated for each model. The Shapley additive explanations (SHAP) values are also calculated to identify the significant variables in GBDT.

Results

There were 8792 patients who underwent emergency colorectal surgery. As a result, the AUROC values of 0.742, 0.782, 0.814, and 0.768 were obtained for logistic regression, random forests, GBDT, and MLP. According to SHAP values, age, colorectal cancer, use of laparoscopy, and some laboratory variables, including serum lactate dehydrogenase serum albumin, and blood urea nitrogen, were significantly associated with in-hospital mortality.

Conclusion

We successfully generated a machine-learning prediction model, including GBDT, with the best prediction performance and exploited the potential for use in evaluating in-hospital mortality risk for patients who undergo emergency colorectal surgery.

Estimating postoperative mortality in colorectal surgery- a systematic review of risk prediction models

PURPOSE

Risk prediction models are frequently used to support decision-making in colorectal surgery but can be inaccurate. Machine learning (ML) is becoming increasingly popular, and its application may increase predictive accuracy. We compared conventional risk prediction models for postoperative mortality (based on regression analysis) with ML models to determine the benefit of the latter approach.

METHODS

The study was registered in PROSPERO(CRD42022364753). Following the PRISMA guidelines, a systematic search of three databases (MEDLINE, EMBASE, WoS) was conducted (from 1/1/2000 to 29/09/2022). Studies were included if they reported the development of a risk model to estimate short-term postoperative mortality for patients undergoing colorectal surgery. Discrimination and calibration performance metrics were compared. Studies were evaluated against CHARMS and TRIPOD criteria.

RESULTS

3,052 articles were screened, and 45 studies were included. The total sample size was 1,356,058 patients. Six studies used ML techniques for model development. Most studies (n = 42) reported the area under the receiver operating characteristic curve (AUROC) as a measure of discrimination. There was no significant difference in the mean AUROC values between regression models (0.833 s.d. ± 0.52) and ML (0.846 s.d. ± 0.55), p = 0.539. Calibration statistics, which measure the agreement between predicted estimates and observed outcomes, were less consistent. Risk of bias assessment found most concerns in the data handling and analysis domains of eligible studies.

CONCLUSIONS

Our study showed comparable predictive performance between regression and ML methods in colorectal surgery. Integration of ML in colorectal risk prediction is promising but further refinement of the models is required to support routine clinical adoption.

Critical Analysis of the Causes of In-Hospital Mortality following Colorectal Resection: A Queensland Audit of Surgical Mortality (QASM) Registry Study

Background

Colorectal resection is a major gastrointestinal operation. Improvements in peri-operative care has led to improved outcomes; however, mortalities still occur. Using data from the Queensland Audit of Surgical Mortality (QASM), this study examines the demographic and clinical characteristics of patients who died in hospital following colorectal resection, and also reports the primary cause of death in this population.

Methods

Patients who died in hospital following colorectal resection in Queensland between January 2010 and December 2020 were identified from the QASM database.

Results

There were 755 patients who died in the 10 year study period. Pre-operatively, the risk of death as subjectively determined by operating surgeons was ‘considerable’ in 397 cases (53.0%) and ‘expected’ in 90 cases (12.0%). The patients had a mean of 2.7 (±1.5) co-morbidities, and a mean American Society of Anaesthesiologists (ASA) score of 3.6 (±0.8). Operations were categorised as emergency in 579 patients (77.2%), with 637 patients (85.0%) requiring post-operative Intensive Care Unit (ICU) support. The primary cause of death was related to a surgical cause in 395 patients (52.7%) and to a medical cause in 355 patients (47.3%). The primary causes of death were advanced surgical pathology (n=292, 38.9%), complications from surgery (n=103, 13.7%), complications arising from pre-existing medical co-morbidity (n=282, 37.6%) or new medical complications unrelated to pre-existing conditions (n=73, 9.7%).

Conclusions

Patients who died had significant co-morbidities and often presented emergently with an advanced surgical pathology. Surgical and medical causes of death both contributed equally to the mortality burden.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00268-022-06534-9.

Data analytics and artificial intelligence in predicting length of stay, readmission, and mortality: a population-based study of surgical management of colorectal cancer

Data analytics and artificial intelligence (AI) have been used to predict patient outcomes after colorectal cancer surgery. A prospectively maintained colorectal cancer database was used, covering 4336 patients who underwent colorectal cancer surgery between 2003 and 2019. The 47 patient parameters included demographics, peri- and post-operative outcomes, surgical approaches, complications, and mortality. Data analytics were used to compare the importance of each variable and AI prediction models were built for length of stay (LOS), readmission, and mortality. Accuracies of at least 80% have been achieved. The significant predictors of LOS were age, ASA grade, operative time, presence or absence of a stoma, robotic or laparoscopic approach to surgery, and complications. The model with support vector regression (SVR) algorithms predicted the LOS with an accuracy of 83% and mean absolute error (MAE) of 9.69 days. The significant predictors of readmission were age, laparoscopic procedure, stoma performed, preoperative nodal (N) stage, operation time, operation mode, previous surgery type, LOS, and the specific procedure. A BI-LSTM model predicted readmission with 87.5% accuracy, 84% sensitivity, and 90% specificity. The significant predictors of mortality were age, ASA grade, BMI, the formation of a stoma, preoperative TNM staging, neoadjuvant chemotherapy, curative resection, and LOS. Classification predictive modelling predicted three different colorectal cancer mortality measures (overall mortality, and 31- and 91-days mortality) with 80-96% accuracy, 84-93% sensitivity, and 75-100% specificity. A model using all variables performed only slightly better than one that used just the most significant ones.

Impact of the COVID-19 pandemic on the management of colorectal cancer in Denmark

INTRODUCTION

The COVID-19 pandemic has had a global impact on cancer care but the extent to which this has affected the management of colorectal cancer (CRC) in different countries is unknown. CRC management in Denmark was thought to have been relatively less impacted than in other nations during the first wave of the pandemic. The aim of this study was to determine the pandemic's impact on CRC in Denmark.

METHODS

The Danish national cancer registry identified patients with newly diagnosed with CRC from 1 March 2020 to 1 August 2020 (pandemic interval) and corresponding dates in 2019 (prepandemic interval). Data regarding clinicopathological demographics and perioperative outcomes were retrieved and compared between the two cohorts.

RESULTS

Total CRC diagnoses (201 versus 359 per month, P = 0.008) and screening diagnoses (38 versus 80 per month, P = 0.016) were both lower in the pandemic interval. The proportions of patients presenting acutely and the stage at presentation were, however, unaffected. For those patients having surgery, both colonic and rectal cancer operations fell to about half the prepandemic levels: colon (187 (i.q.r. 183-188) to 96 (i.q.r. 94-112) per month, P = 0.032) and rectal cancers (63 (i.q.r. 59-75) to 32 (i.q.r. 28-42) per month, P = 0.008). No difference was seen in surgical practice or postoperative 30-day mortality rate (colon 2.2 versus 2.2 per cent, P = 0.983; rectal 1.0 versus 2.9 per cent, P = 0.118) between the cohorts. Treatment during the pandemic interval was not independently associated with death at 30 or 90 days.

CONCLUSION

The initial wave of the COVID-19 pandemic reduced the number of new diagnoses made and number of operations but had limited impact on technique or outcomes of CRC care in Denmark.

A Prediction Model for Severe Complications after Elective Colorectal Cancer Surgery in Patients of 70 Years and Older

Introduction Older patients have an increased risk of morbidity and mortality after colorectal cancer (CRC) surgery. Existing CRC surgical prediction models have not incorporated geriatric predictors, limiting applicability for preoperative decision-making. The objective was to develop and internally validate a predictive model based on preoperative predictors, including geriatric characteristics, for severe postoperative complications after elective surgery for stage I-III CRC in patients ≥70 years.

PATIENTS AND METHODS

A prospectively collected database contained 1088 consecutive patients from five Dutch hospitals (2014-2017) with 171 severe complications (16%). The least absolute shrinkage and selection operator (LASSO) method was used for predictor selection and prediction model building. Internal validation was done using bootstrapping.

RESULTS

A geriatric model that included gender, previous DVT or pulmonary embolism, COPD/asthma/emphysema, rectal cancer, the use of a mobility aid, ADL assistance, previous delirium and polypharmacy showed satisfactory discrimination with an AUC of 0.69 (95% CI 0.73-0.64); the AUC for the optimism corrected model was 0.65. Based on these predictors, the eight-item colorectal geriatric model (GerCRC) was developed.

CONCLUSION

The GerCRC is the first prediction model specifically developed for older patients expected to undergo CRC surgery. Combining tumour- and patient-specific predictors, including geriatric predictors, improves outcome prediction in the heterogeneous older population.

Nomograms for morbidity and mortality after oncologic colon resection in the enhanced recovery era: results from a multicentric prospective national study

PURPOSE

Predicting postoperative complications and mortality is important to plan the surgical strategy. Different scores have been proposed before to predict them but none of them have been yet implemented into the routine clinical practice because their difficulties and low accuracy with new surgical strategies and enhanced recovery. The main aim of this study is to identify risk factors for postoperative morbidity and mortality after colonic resection (CR) without protective stomas, in order to develop a comprehensive, up-to-date, simple, reliable, and applicable model for the preoperative assessment of patients with colon cancer.

METHODS

Multivariable analysis was performed to identify risk factors for 60-day morbidity and mortality. Coefficients derived from the regression model were used in the nomograms to predict morbidity and mortality.

RESULTS

Three thousand one hundred ninety-three patients from 52 hospitals were included into the analysis. Sixty-day postoperative complications rate was 28.3% and the mortality rate was 3%. In multivariable analysis the independent risk factors for postoperative complications were age, male gender, liver and pulmonary diseases, obesity, preoperative albumin, anticoagulant treatment, open surgery, intraoperative complications, and urgent surgery. Independent risk factors for mortality were age, preoperative albumin anticoagulant treatment, and intraoperative complications.

CONCLUSIONS

Risk factors for morbidity and mortality after CR for cancer were identified and two easy predictive tools were developed. Both of them could provide important information for preoperative consultation and surgical planning in the time of enhance recovery.

Robotic colorectal surgery in Australia: a cohort study examining clinical outcomes and cost

Objective The aim of this study was to compare robotic versus laparoscopic colorectal operations for clinical outcomes, safety and cost. Methods A retrospective cohort study was performed of 213 elective colorectal operations (59 robotic, 154 laparoscopic), matched by surgeon and operation type. Results No differences in age, body mass index, median American Society of Anesthesiologists score or presence of cancer were observed between the laparoscopic or robotic surgery groups. However, patients undergoing robotic colorectal surgery were more frequently male (P = 0.004) with earlier T stage tumours (P = 0.02) if cancer present. Procedures took longer in cases of robotic surgery (302 vs 130 min; P < 0.001), and patients in this group were more frequently admitted to intensive care units (P < 0.001). Overall length of stay was longer (7 vs 5 days; P = 0.03) and consumable cost was A$2728 higher per patient in the robotic surgery group. Conclusion Robotic colorectal surgery appears to be safe compared with current laparoscopic techniques, albeit with longer procedure times and overall length of stay, more frequent intensive care admissions and higher consumables cost. What is known about the topic? Robotic surgery is an emerging alternative to traditional laparoscopic approaches in colorectal surgery. International trials suggest the two techniques are equivalent in safety. What does this paper add? This is an original cohort study examining clinical outcomes in Australian colorectal robotic surgery. The data suggest it may be safe, but this paper demonstrates key issues in the implementation and audit of novel surgical technologies in relatively low-volume centres. What are implications for practitioners? In our study, patients undergoing robotic colorectal surgery at a single centre in Australia had equivalent measured clinical outcomes to those undergoing laparoscopic surgery. However, practitioners may counsel patients that robotic procedures are typically longer and more expensive, with a longer overall hospital admission and a higher likelihood of intensive care admission.

Special needs of frail people undergoing emergency laparotomy surgery

There are now over 30 000 emergency laparotomies under taken in the UK every year, a figure that is increasing year on year. Over half of these people are aged over 70 years old. Frailty is commonly seen in this population and becomes increasingly common with age and is seen in over 50% of elderly emergency laparotomies in people aged over 85 years old. In older people who undergo surgery one third will have died within one year of surgery, a figure which is worse in frail individuals. For those that do survive, post-operative morbidity is worse and 30% of frail older people do not return to their own home. In the UK, the National Emergency Laparotomy Audit (NELA) is leading the way in providing the evidence base in this population group. Beyond collecting data on every Emergency Laparotomy undertaken in the UK, it is also key in driving improvement in care. Their most recent report highlights that only 23% of patients over 70 years received geriatric involvement following surgery. More encouragingly, the degree of multidisciplinary geriatric involvement seems to be increasing. In the research setting, well designed studies focusing on the older frail emergency laparotomy patient are underway. It is anticipated that these studies will better define outcomes following surgery, improving the communication and decision making between patients, relatives, carers and their surgical teams.

Risk prediction models for postoperative outcomes of colorectal cancer surgery in the older population - a systematic review

BACKGROUND

An increasing number of patients with Colorectal Cancer (CRC) is 65 years or older. We aimed to systematically review existing clinical prediction models for postoperative outcomes of CRC surgery, study their performance in older patients and assess their potential for preoperative decision making.

METHODS

A systematic search in Pubmed and Embase for original studies of clinical prediction models for outcomes of CRC surgery. Bias and relevance for preoperative decision making with older patients were assessed using the CHARMS guidelines.

RESULTS

26 prediction models from 25 publications were included. The average age of included patients ranged from 61 to 76. Two models were exclusively developed for 65 and older. Common outcomes were mortality (n = 10), anastomotic leakage (n = 7) and surgical site infections (n = 3). No prediction models for quality of life or physical functioning were identified. Age, gender and ASA score were common predictors; 12 studies included intraoperative predictors. For the majority of the models, bias for model development and performance was considered moderate to high.

CONCLUSIONS

Prediction models are available that address mortality and surgical complications after CRC surgery. Most models suffer from methodological limitations, and their performance for older patients is uncertain. Models that contain intraoperative predictors are of limited use for preoperative decision making. Future research should address the predictive value of geriatric characteristics to improve the performance of prediction models for older patients.

Risk factors for leak, complications and mortality after ileocolic anastomosis: comparison of two anastomotic techniques

INTRODUCTION

There are no definitive data concerning the ideal configuration of ileocolic anastomosis. Aim of this study was to identify perioperative risk factors for anastomotic leak and for 60-day morbidity and mortality after ileocolic anastomoses (stapled vs handsewn).

MATERIALS AND METHODS

This is a STROBE-compliant study. Demographic and surgical data were gathered from patients with an ileocolic anastomosis performed between November 2010 and September 2016 at a tertiary hospital. Anastomoses were performed using standardised techniques. Independent risk factors for anastomotic leak, complications and mortality were assessed.

RESULTS

We included 477 patients: 53.7% of the anastomoses were hand sewn and 46.3% stapled. Laterolateral anastomosis was the most common configuration (93.3%). Anastomotic leak was diagnosed in 8.8% of patients and 36 were classified as major anastomotic leak (7.5%). In the multivariate analysis, male sex (P = 0.014, odds ratio, OR, 2.9), arterial hypertension (P = 0.048, OR 2.29) and perioperative transfusions (P < 0.001, OR 2.4 per litre) were independent risk factors for major anastomotic leak. The overall 60-day complication rate was 27.3%. Male sex (31.3% vs female 22.3%, P = 0.02, OR 1.7), diabetes (P = 0.03 OR 2.0), smoking habit (P = 0.04, OR 1.8) and perioperative transfusions (P < 0.001, OR 3.3 per litre) were independent risk factors for postoperative morbidity. The 60-day-mortality rate was 3.1% and no significant risk factors were identified.

CONCLUSION

Anastomotic leak after ileocolic anastomosis is a relevant problem. Male sex, arterial hypertension and perioperative transfusions were associated with major anastomotic leak. Conversion to open surgery was more frequently associated with perioperative death.

National cohort study on postoperative risks after surgery for submucosal invasive colorectal cancer

Background

The decision to perform surgery for patients with T1 colorectal cancer hinges on the estimated risk of lymph node metastasis, residual tumour and risks of surgery. The aim of this observational study was to compare surgical outcomes for T1 colorectal cancer with those for more advanced colorectal cancer.

Methods

This was a population‐based cohort study of patients treated surgically for pT1–3 colorectal cancer between 2009 and 2016, using data from the Dutch ColoRectal Audit. Postoperative complications (overall, surgical, severe complications and mortality) were compared using multivariable logistic regression. A risk stratification table was developed based on factors independently associated with severe complications (reintervention and/or mortality) after elective surgery.

Results

Of 39 813 patients, 5170 had pT1 colorectal cancer. No statistically significant differences were observed between patients with pT1 and pT2–3 disease in the rate of severe complications (8·3 versus 9·5 per cent respectively; odds ratio (OR) 0·89, 95 per cent c.i. 0·80 to 1·01, P = 0·061), surgical complications (12·6 versus 13·5 per cent; OR 0·93, 0·84 to 1·02, P = 0·119) or mortality (1·7 versus 2·5 per cent; OR 0·94, 0·74 to 1·19, P = 0·604). Male sex, higher ASA grade, previous abdominal surgery, open approach and type of procedure were associated with a higher severe complication rate in patients with pT1 colorectal cancer.

Conclusion

Elective bowel resection was associated with similar morbidity and mortality rates in patients with pT1 and those with pT2–3 colorectal carcinoma.

CCR-CARESS score for predicting operative mortality in patients with colorectal cancer

BACKGROUND

The aim of this study was to assess factors associated with outcomes after surgery for colorectal cancer and to design and internally validate a simple score for predicting perioperative mortality.

METHODS

Patients undergoing surgery for primary invasive colorectal cancer in 22 centres in Spain between June 2010 and December 2012 were included. Clinical variables up to 30 days were collected prospectively. Multiple logistic regression techniques were applied and a risk score was developed. The Hosmer-Lemeshow test was applied and the area under the receiver operating characteristic (ROC) curve (AUC, with 95 per cent c.i.) was estimated.

RESULTS

A total of 2749 patients with a median age of 68·5 (range 24-97) years were included; the male : female ratio was approximately 2 : 1. Stage III tumours were diagnosed in 32·6 per cent and stage IV in 9·5 per cent. Open surgery was used in 39·3 per cent, and 3·6 per cent of interventions were urgent. Complications were most commonly infectious or surgical, and 25·5 per cent of patients had a transfusion during the hospital stay. The 30-day postoperative mortality rate was 1·9 (95 per cent c.i. 1·4 to 2·4) per cent. Predictive factors independently associated with mortality were: age 80 years or above (odds ratio (OR) 2·76), chronic obstructive pulmonary disease (COPD) (OR 3·62) and palliative surgery (OR 10·46). According to the categorical risk score, a patient aged 80 years or more, with COPD, and who underwent palliative surgery would have a 23·5 per cent risk of death within 30 days of the intervention.

CONCLUSION

Elderly patients with co-morbidity and palliative intention of surgery have an unacceptably high risk of death.

ACS-NSQIP risk calculator predicts cohort but not individual risk of complication following colorectal resection

OBJECTIVE

Compare the ACS-NSQIP risk calculator with institutional risk for colorectal surgery.

METHODS

Actual and predicted outcomes were compared for both cohort and individuals.

RESULTS

For the cohort, the risk calculator was accurate for 7/8 outcomes; there were more serious complications than predicted (19.4 vs 14.7%, p < 0.05). Risk calculator Brier scores and null Brier scores were comparable.

PATIENTS

with better outcomes than predicted were current smokers (OR 4.3 95% CI 1.2-15.4), ASA ≥ 3 (OR 10.4, 95% CI 2.8-39.2), underwent total/subtotal colectomy (OR 3.5, 95% CI 1.1-12.2) or operated by Surgeon 2 (OR 2.9, 95% CI 1.4-11.6). Patients with serious complications who had low predicted risk had low ASA (OR 10.5, 95% CI 1.3-82.6), and underwent operation by Surgeon 2 (OR 11.8, 95% CI 2.5, 55.2).

LIMITATIONS

Single center study, sample size may bias subgroup analyses.

CONCLUSIONS

The ACS NSQIP calculator did not predict outcome better than sample risk.