Machine Learning for Optimal Individual Survival Prediction in Resectable Upper Gastrointestinal Cancer

doi:10.21203/rs.3.rs-1318132/v1

Purpose

Surgical oncologists are frequently confronted with the question of expected long-term prognosis. The aim of this study was to apply machine learning algorithms to optimize survival prediction after oncological resection of gastroesophageal cancers.

Methods

Eligible patients underwent oncological resection of gastric or distal esophageal cancer between 2001 and 2020 at Heidelberg University Hospital, Department of General Surgery. Machine learning methods such as multi-task logistic regression and survival forests were compared with usual algorithms to establish an individual estimation.

Results

The study included 117 variables with a total of 1,360 patients. The overall missingness was 1.3%. Out of eight machine learning algorithms, the random survival forest (RSF) performed best with a concordance-index of 0.736 and an integrated Brier score of 0.166. The RSF demonstrated a mean area under the curve (AUC) of 0.814 over a time period of 10 years after diagnosis. The most important long-term outcome predictor was lymph node ratio with a mean AUC of 0.730. A numeric risk score was calculated by the RSF for each patient and three risk groups were defined accordingly. Median survival time was 18.8 months in the high-risk group, 44.6 months in the medium-risk group and above 10 years in the low-risk group.

Conclusion

The results of this study suggest that RSF is most appropriate to accurately answer the question of long-term prognosis. Furthermore, we could establish a compact risk score model with 20 input parameters and thus provide a clinical tool to improve prediction of oncological outcome after upper gastrointestinal surgery.

gastric cancer

esophageal cancer

machine learning

survival analysis

oncological outcome

The demand for adequate prediction of oncological outcome after upper gastrointestinal surgery will increase due to a rising global incidence of esophageal and gastric carcinomas (Malhotra et al. 2017). In this regard, patient profiles are heterogenous due to varying in-hospital courses and additional information about clinical and histopathological characteristics which make the individual prognosis difficult to predict. While most patient characteristics are hardly improvable and are rather predetermined by the nature of the disease, there might be critical periods during the treatment of a malignancy such as the phase of oncological resection when many fundamental conditions for long-term outcome are set. This way, the in-hospital stay for surgery is literally an incisive event in the clinical course of one individual patient. However, surgical morbidity can significantly influence the overall and long-term survival in gastric cancer (Kulig et al. 2021). Eventually, clinicians are confronted with a patient curious to know what the information gathered so far from surgery actually means for his or her individual prognosis. While the patient is then frequently referred to the oncologist who will further supervise the prospective course, there may be a more concrete estimation to offer at the time of discharge. Surgeons with an oncologic focus need the tools to provide these answers to adequately care for their patients.

The motivation to accurately predict the individual prognosis of one patient also derives from the growing interest to assess individualized patient profiles in times of personalized medicine (Hu and Steingrimsson 2018). Due to the complexity of each patient case, differences in individual prognosis could eventually implicate different (adjuvant) therapeutic regimens. In this context, the vast amount of clinical data should be viewed as another component of “omics”-based technology and as another step towards personalized treatment.

Machine learning techniques can help to determine significant associations based on complex and vast data (Kourou et al. 2015). Retrospective machine learning studies have already delivered suggestive data that artificial intelligence may be promising in case of multiple potential factors with unclear relations to each other. Since the proposal of the proportional hazards model by David Cox in 1972 (Cox 1972) there were many advances to improve survival prediction, especially since computational possibilities have increased. Among those, machine learning algorithms have recently shown promising results (Zhu et al. 2020). According to the current literature there are many works regarding the automated prediction of time-dependent, censored data with machine learning methods. Besides survival being the most typical and classic scenario, there have also been approaches to predict heart failure (Panahiazar et al. 2015), kidney transplant durability (Sekercioglu et al. 2021) and many other settings. Spooner et al. were able to demonstrate that machine learning algorithms for survival analysis were applicable to the clinical task of predicting dementia (Spooner et al. 2020). Based on two separate datasets, it was possible to reach a concordance index (c-index) of 0.82 and 0.93, respectively.

However, the prediction of oncological outcome after diagnosis of a malignant disease is still the most common target variable in machine learning studies. To this date, there are only few medical publications in the current literature dealing with machine learning models for survival analysis of oncologically resected upper gastrointestinal cancers. Akcay et al. have analyzed gastric cancer patients after chemoradiation with a comparable methodology (Akcay et al. 2020) and tested a random forest algorithm to predict tumor relapse in terms of distant metastases or peritoneal recurrence. Compared to other algorithms such as logistic regression, multilayer perceptron and extreme gradient boosting the random forest scored highest with an area under the curve (AUC) of 0.97 for peritoneal recurrence. However, the random forest algorithm did not show convincing results for overall survival with an AUC of only 0.59 and the authors did not apply a random survival forest methodology. Jiang et al. have utilized a LASSO cox analysis to predict oncological outcome of gastric cancer patients analyzing the radiomic signature of PET computer tomography (Jiang et al. 2018). It is notable that the authors evaluated imaging data and reached a powerful prediction for disease-free survival and overall survival with a c-index of 0.786. From a methodological point of view, Pölsterl et al. have found that feature extraction in the setting of survival prediction is not properly applicable in case of small sample sizes such as approximately 500 patients (Pölsterl et al. 2016). On the other hand, in case of larger sample sizes with more than 2500 individuals feature extraction performs similarly to feature selection methods. Due to this reason the effect of feature extraction is supposedly maximized in-between the mentioned case numbers.

To our knowledge, the topic of survival prediction in case of oncologically resected upper gastrointestinal cancer is not sufficiently covered by machine learning methods. However, the opportunity of pre- and postoperatively supervising an individual patient course offers an ideal setting to predict long-term outcome based on the collected data. The aim of this study was therefore to apply various machine learning algorithms to optimize survival prediction after oncological resection of gastroesophageal cancer compared to previously established methods.

Data collection and follow-up

Eligible subjects of this study were patients with distal esophageal, gastroesophageal junction or gastric cancer who underwent oncological resection between September 2001 and December 2020 at Heidelberg University Hospital, Department of General Surgery. All patients provided written consent for data collection and analysis. The data was initially collected prospectively in a clinical database. The trial protocol was approved by the ethics committee at the University of Heidelberg (committee’s approval: S-635/2013) and was performed in accordance with the Declaration of Helsinki, Good Clinical Practices as well as local ethics and legal requirements. To acquire long-term survival data, all patients were systematically followed-up via continuous surveys. The relevant survival data for this patient collective was gathered until October 2020.

Inclusion and exclusion criteria

Out of the main collective, only those patients were further taken into account with a preoperatively diagnosed and histologically proven adenocarcinoma of the distal esophagus, the esophagogastric junction or stomach. Patients were excluded who had squamous cell carcinoma at any location. Also, only radical oncological resections were evaluated as opposed to exploratory laparotomies with palliative treatment. Furthermore, the type of oncological resection was limited to (sub-)total gastrectomy, gastrectomy with transhiatal extension, proximal gastrectomy, abdominothoracic esophagectomy and discontinuity resection of the esophagus. Variables with a missingness greater than 50% were generally excluded from further analysis. Also, patients were excluded with a case-specific missingness of more than 10%.

Data structure

Eventually, the dataset had a total of 117 features and 1,360 patients. Out of the whole dataset, 92 variables were categorical (with 77 Boolean variables) and 25 variables were numeric. Table 1 gives a short overview of the parameters including all dependent variables (or end points, respectively). There were obvious survival predictors and other variables which had to be excluded due to collinearity (see Supplementary Figure 1). Regarding survival data, 53.5% of the records were censored with an equal distribution over the whole timespan. Supplementary Figure 2 demonstrates the censored data added on top of the patients that verifiably passed away and plotted against the follow-up period.

Missingness and imputation

After the application of the above-mentioned criteria, there was a remaining missingness in 2,033 datapoints and thus an overall missingness of 1.3% without any duplicate values (see Supplementary Figure 3). For imputation, missingness at random was assumed and multiple imputations with n = 1000 iterations via IterativeImputer from the Sci-kit learn package (Pedregosa et al. 2011) were performed on Python 3.9 (Van Rossum G and Drake FL 2009).

Statistical analysis

All statistical analyses were performed on Python 3.9 with packages such as scikit-learn 0.24.2 by Pedregosa et al. (Pedregosa et al. 2011). To compare various machine learning algorithms the package PySurvival by Fotso et al. (Stephane Fotso 2019) was implemented. The performances of the different algorithms were scored according to the c-index (Uno et al. 2011). The inverse probability of censoring weights (IPCW c-index) was considered as an alternative to the standard c-index which is independent of the distribution of censored cases in the test data. This was relevant due to the comparably high prevalence of censored time points in the present dataset (see Supplementary Figure 2). To establish another performance score, the integrated Brier-score (further abbreviated as IBS) was utilized (Steyerberg et al. 2010). The IBS was eventually plotted to demonstrate prediction error with a cut-off limit of 0.25 considered as critical. Furthermore, the actual survival function was plotted against the predicted function by the model. Finally, time-dependent evaluation of the area under the curve (AUC) was performed for selected machine learning ensembles provided by the scikit-survival package version 0.15.1 (Pölsterl 2020).

Feature selection

We did not perform feature selection before application of algorithms since previous works by Spooner et al. have shown that feature selection in this scenario does not significantly improve model performance (Spooner et al. 2020). The authors have applied seven different feature selection algorithms before running 5-fold cross validation with 5 repeats. The results did not show any significant improvement of test statistics and most machine learning algorithms have a feature selection method internalized already. However, we identifiedal the 20 most relevant predictors after fitting the corresponding machine learning algorithm via permutation-based importance evaluation provided by ELI5 (Arya et al. 2020).

Machine learning algorithms

The classic Cox proportional hazards model introduced by David Cox in 1972 is probably the most popular survival prediction model with an easy to interpret statistic (Cox 1972). We also tested a non-linear Cox proportional hazards model (also called DeepSurv) which is a multi-layer perceptron (Katzman et al. 2018). However, the authors’ work is also known for the focus on recommender functions and treatment decision making. The Linear Multi-Task Logistic Regression (L-MTLR) was introduced by Yu et al. in 2011 and operates on the basis of multiple logistic regressions (Yu et al. 2011). It may be considered as another general alternative to the Cox regression model. To increase flexibility, the Neural Multi-Task Logistic Regression (N-MTLR) was introduced in 2019 (Stephane Fotso 2019). To fully represent all available survival prediction models, we also included one parametric model, the Gompertz model, as the relatively strongest test statistic compared to the Weibull and Exponential model which are not demonstrated in this work. Last but not least, several random survival forest models were analyzed including the classic Random Survival Forest (RSF) by Ishwaran et al. (Ishwaran et al. 2008). The Conditional Survival Forest model was developed by Wright et al. to improve splitting of the RSF by applying maximally selected rank statistics for the split point selection (Wright et al. 2017). The Extra Trees incorporated in PySurvival are an extension of the Extremely Randomized Trees (Geurts et al. 2006) which is a supervised learning method with almost totally randomized decision trees. The model is eventually independent of the output values of the learning sample and impresses with its high computational efficiency.

Hyperparameter optimization

We performed hyperparameter optimization through defining the closest neighbor parameter values for each algorithm. Thereafter, the optimization was executed via Halving Grid Search included in the latest version of scikit-learn (Pedregosa et al. 2011). Halving Grid Search operates via successive halving of the candidate hyperparameters and their effect on the final test statistic as measured by the c-index. Although Halving Grid Search was not yet applied in many scientific works, it has already been shown that it yields equivalent results for hyperparameter tuning compared to usual brute-force Grid Searching, however, with a much higher computation efficiency (Sraitih et al. 2021).

The study included 1,360 patients with oncological resection between 2001 and 2020 and 117 variables. Table 2 summarizes all patient characteristics which were mainly evaluated as potentially relevant predictors for the machine learning models except for obvious outcome parameters.

First of all, we tested the introduced machine learning algorithms which are demonstrated with the according test statistic in Figure 1. The standard Cox proportional hazards (CPH) model is demonstrated in Figure 1a and reached a c-index of 0.645 and an integrated Brier score (or IBS) of 0.221. The prediction error curve is depicted in the last column of Figure 1 with the values for the Brier score as an integral. The non-linear Cox proportional hazards model (see Figure 1b) was able to improve the statistic with a c-index of 0.681 and an IBS of 0.194. Note that while both Cox proportional hazards models initially have a comparably good score, the integral reaches the critical limit of 0.25 as the timespan reaches 10 years after cancer diagnosis.

Compared to the first two calculations, the c-index could not be improved by linear multi-task logistic regression (see Figure 1c, c-index = 0.673, IBS = 0.229). However, the neural multi-task logistic regression reached a clearly better prediction of the actual survival function with a root mean squared error (RMSE) of the actual predicted survival curve of 9.188 (see Figure 1d, c-index = 0.672, IBS = 0.254). Note that both regression methods also exceed the IBS cut-off value of 0.25 which is generally considered as an acceptable limit. The Gompertz model as an example for a parametric model was not able to significantly improve the test statistic and showed relatively weak results (see Figure 1e, c-index = 0.677, IBS = 0.194).

However, the CPH models could be outperformed by all three survival forest methods with the RSF being the strongest prediction model (see Figure 1h, c-index = 0.736, IBS = 0.166). The Extra Survival Trees (see Figure 1g, c-index = 0.736, IBS = 0.167) and the Conditional Survival Trees (see Figure 1f, c-index = 0.726, IBS = 0.166) showed similarly strong results which could still not outperform the RSF even after hyperparameter optimization. The accurate prediction by RSF is again underlined by the direct comparison between actual and predicted survival function demonstrated in the second column. Here, the root-mean-square error (RMSE) of the RSF was calculated as 6.224.

To establish a more practical and compact approach, we selected the most important features identified by the RSF. Table 3 shows the permutation-based importance of the 20 most important parameters identified by the RSF model. While the individual weights of the predictors are not remarkably high (except for lymph node ratio), the RSF model can still build the prediction based on all variables of the dataset. The weights of the predictors are to be interpreted in a fashion that for instance the omittance of lymph node ratio in the model would evoke a change of the resulting c-index by 0.118 with the specified 95% confidence range.

The time-dependent area under the curve (AUC) is separately demonstrated for the six most important predictors. It is remarkable that the AUC factors such as duration of intensive care, postoperative complications and intraoperative blood loss lose their predictive value rapidly as time passes. In contrast, the significance of lymph node ratio increases postoperatively and stays stable on an AUC level above 0.65 more than 5 years after cancer diagnosis (see Figure 2a).

While predictors have a time-dependent importance, the prediction models also show a time dependent accuracy. Of note, the RSF algorithm outperforms the CPH model on the time-dependent scale with an AUC of 0.821 (as opposed to 0.720 for the CPH model, see Figure 2b). The remaining machine learning models also showed better time-dependent performances than CPH but were not able to outperform the RSF algorithm. While all models were able to predict survival with a very high score above 0.9 in the very first months, the predictions generally tended to decrease in accuracy while time progressed. However, the long-term survival prediction was most successful according to the RSF model.

Figure 3a shows a risk scoring model based on the test statistic calculated by the RSF algorithm. A numeric risk score is assigned to each patient ranging from 4.7 to 7.1. Three different colors were utilized to depict the low-, medium- and high-risk group. The differentiation of three groups was performed manually based on the distribution of risk scores. Figure 3b shows the survival curves of all individuals classified by the scoring system as low-, medium- or high-risk within the same predefined test group. The three survival curves differ significantly according to the log-rank test (p < 0.0001) with the low-risk group having a 5-year survival rate of 73.18%, the medium-group showing 45.39% and the high-risk group finally 14.87%. Median survival time was 18.754 months in the high-risk group, 44.557 months in the medium-risk group and incalculable in the low-risk group since the survival rate did not fall below 50% in the observed long-term time interval of 10 years. Also, the survival curves can be separated into more groups to enable a further stratification of risk groups (not shown).

Finally, the 20 most relevant predictors from the permutation-based importance scoring (see Table 3) were selected to establish a more compact RSF model. Supplementary Figure 4 shows the distribution of the risk scores resulting from the compact RSF model. A risk score between 0 and 5.3 was considered as low-risk, a score between 5.3 and 6.1 as medium-risk and a score above 6.1 as high-risk for decease after resection. The according survival curves of the three different groups from the test cohort are demonstrated in Supplementary Figure 5. The low-risk group had an incalculable median survival longer than 10 years, the medium-risk group had a median survival of 85.639 months and the high-risk group 20.721 months.

To improve the prediction of overall survival after gastroesophageal cancer resections we tested eight machine learning algorithms in this retrospective survival analysis to determine the best prediction model for oncological outcome. The Cox proportional hazards (CPH) model is by far the most frequently applied model to evaluate significant survival predictors with coefficient metrics that are easy to interpret. However, the CPH model relies on a manual selection of the most important features with potential omittance of important features. Therefore, the CPH model may not be able to handle complex and vast data with multidimensional features. Machine learning methods have shown since their application in econometrics and biosciences that it is especially useful for prediction problems (Jordan and Mitchell 2015). Using the c-index as a measure of performance, the random survival forest (RSF) model achieved the most accurate prediction with a c-index of 0.736. This value is generally considered in the setting of time-dependent censored data such as long-term survival as a good to strong prediction model. Moreover, we could show that the RSF algorithm was able to outperform the CPH model as well as all other machine learning models. In our setting, the RSF is robust which is proofed by 10-fold cross-validation. We therefore believe that the RSF model is the most suitable machine learning algorithm to predict oncological outcome after curative resection of gastroesophageal adenocarcinoma.

The strength as well as the limitation of this study is the availability of detailed and high-resolution data for each patient. On the one hand, it was possible to establish an excellent prediction model under these circumstances. However, it might not be feasible in every setting to extract all presented data for each patient who is treated in a surgical hospital. To establish a more practical approach, we selected the 20 most important features identified by the RSF. By reducing the necessary input parameters from 117 to 20 and thus avoiding any tedious or unnecessary data management, it may be possible to make oncological predictions more practicable for the surgical oncologist. The resulting test statistic of this compact RSF model performs in a comparable strength and could be applied in the future. In both cases, extended or compact RSF, the test statistic outperformed the Cox proportional hazards model. It is also imaginable to establish an online application similar to the surgical risk calculator by the American College of Surgeons (Bilimoria et al. 2013) enabling clinicians to enter the 20 variables anonymously since no identifying information is needed to handle the algorithm. The application can then generate an individual assessment based on the information that is present at the time of discharge. Given the anonymous information, it would also be possible to show specific survival curves according to the trained model and the risk group that the patient belongs to. This is statistically legitimate since the plotted survival curve represents the survival function for all patients with the same risk score range over the time span of 10 years.

The application of an individual and numeric risk score summarizes all available clinical and histopathological data that is predictive for long-term oncological outcome. Future clinical trials could rely on this risk score to find differences for overall survival which is the actual outcome parameter of primary interest. Thus, it is imaginable that subgroups could be treated with or without adjuvant therapy not according to single features such as TNM-status but based on the affiliation to the risk groups presented in this study. Likewise, high-risk patients with worst prognosis could be analyzed separately to eventually achieve improvements in surveillance and treatment. In this context, it is urgent to identify and characterize high-risk patients with unfavorable histopathological and postoperative parameters who still show a comparably good survival. To this date, it is not properly understood if the mere fact that the tumor was resected may nevertheless have an impact on overall survival for this specific patient collective and it is debatable if immunological reactions play a role. It is also necessary to implement a generally accepted risk score such as the presented machine learning model regardless of resectional status to offer a prognostic tool also for palliative situations.

This study is outstanding due to a systematic follow-up, high data quality and a thorough, state-of-the-art imputation with application of documented Python packages. All reported machine learning models are retrievable and reproducible via open source and have been cited accordingly. However, the clinical relevance of the reported RSF model as well as the other machine learning models needs to be clarified in prospective analyses with inevitably necessary external validation. Furthermore, the value differences between the several survival forest models are limited to some hundredths of the c-index and the discussion of these differences may be of limited clinical relevance. Nevertheless, we could show in a supervised learning setting that the algorithms were able to automatically select and extract predictors that are verifiably relevant for long-term prognosis.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Jin-On Jung, Nerma Crnovrsanin, Naita Maren Wirsik, Henrik Nienhüser, Leila Peters and Thomas Schmidt. The first draft of the manuscript was written by Jin-On Jung and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Ethics Approval

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. The trial protocol was approved by the ethics committee at the University of Heidelberg (committee’s approval: S-635/2013).

Consenct to participate

Informed consent was obtained from all individual participants included in the study.

Akcay M, Etiz D, Celik O, 2020. Prediction of Survival and Recurrence Patterns by Machine Learning in Gastric Cancer Cases Undergoing Radiation Therapy and Chemotherapy. Adv. Radiat. Oncol. 5, 1179–1187. https://doi.org/10.1016/j.adro.2020.07.007
Arya V, Bellamy RKE, Chen P-Y, Dhurandhar A, Hind M, Hoffman SC, Houde S, Liao QV, Luss R, Mourad S, Pedemonte P, Raghavendra R, Richards JT, Sattigeri P, Shanmugam K, Singh M, Varshney KR, Wei D, Zhang Y, 2020. AI Explainability 360: An Extensible Toolkit for Understanding Data and Machine Learning Models. J. Mach. Learn. Res. 21, 1–6.
Bilimoria KY, Liu Y, Paruch JL, Zhou L, Kmiecik TE, Ko CY, Cohen ME, 2013. Development and evaluation of the universal ACS NSQIP surgical risk calculator: A decision aid and informed consent tool for patients and surgeons. J. Am. Coll. Surg. 217, 833-842.e3. https://doi.org/10.1016/j.jamcollsurg.2013.07.385
Cox DR, 1972. Regression Models and Life-Tables. J. R. Stat. Soc. Ser. B 34, 187–202. https://doi.org/10.1111/j.2517-6161.1972.tb00899.x
Geurts P, Ernst D, Wehenkel L, 2006. Extremely randomized trees. Mach. Learn. 63, 3–42. https://doi.org/10.1007/s10994-006-6226-1
Hu C, Steingrimsson JA, 2018. Personalized Risk Prediction in Clinical Oncology Research: Applications and Practical Issues Using Survival Trees and Random Forests. J. Biopharm. Stat. 28, 333–349. https://doi.org/10.1080/10543406.2017.1377730
Ishwaran H, Kogalur UB, Blackstone EH, Lauer MS, 2008. Random survival forests. Ann. Appl. Stat. 2, 841–860. https://doi.org/10.1214/08-AOAS169
Jiang Y, Yuan Q, Lv W, Xi S, Huang W, Sun Z, Chen H, Zhao L, Liu W, Hu Y, Lu L, Ma J, Li T, Yu J, Wang Q, Li G, 2018. Radiomic signature of 18F fluorodeoxyglucose PET/CT for prediction of gastric cancer survival and chemotherapeutic benefits. Theranostics 8, 5915–5928. https://doi.org/10.7150/thno.28018
Jordan MI, Mitchell TM, 2015. Machine learning: Trends, perspectives, and prospects. Science (80-. ). https://doi.org/10.1126/science.aaa8415
Katzman JL, Shaham U, Cloninger A, Bates J, Jiang T, Kluger Y, 2018. DeepSurv: Personalized treatment recommender system using a Cox proportional hazards deep neural network. BMC Med. Res. Methodol. 18. https://doi.org/10.1186/s12874-018-0482-1
Kourou K, Exarchos TP, Exarchos KP, Karamouzis M V., Fotiadis DI, 2015. Machine learning applications in cancer prognosis and prediction. Comput. Struct. Biotechnol. J. 13, 8–17. https://doi.org/10.1016/J.CSBJ.2014.11.005
Kulig P, Nowakowski P, Sierzȩga M, Pach R, Majewska O, Markiewicz A, Kołodziejczyk P, Kulig J, Richter P, 2021. Analysis of prognostic factors affecting short-term and long-term outcomes of gastric cancer resection. Anticancer Res. https://doi.org/10.21873/anticanres.15140
Malhotra GK, Yanala U, Ravipati A, Follet M, Vijayakumar M, Are C, 2017. Global trends in esophageal cancer. J. Surg. Oncol. 115, 564–579. https://doi.org/10.1002/jso.24592
Panahiazar M, Taslimitehrani V, Pereira N, Pathak J, 2015. Using EHRs and Machine Learning for Heart Failure Survival Analysis, in: Studies in Health Technology and Informatics. IOS Press, pp. 40–44. https://doi.org/10.3233/978-1-61499-564-7-40
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay É, Pedregosa, F. and Varoquaux G, Gramfort, A. and Michel V, Thirion, B. and Grisel O, Blondel, M. and Prettenhofer P, Weiss, R. and Dubourg V, Vanderplas, J. and Passos A, Cournapeau, D. and Brucher M, Perrot M, Duchesnay E, 2011. Scikit-learn: Machine Learning in Python. J. Mach. Learn. Res. 12, 2825–2830.
Pölsterl S, 2020. Scikit-survival: A library for time-to-event analysis built on top of scikit-learn. J. Mach. Learn. Res. 21, 1–6.
Pölsterl S, Conjeti S, Navab N, Katouzian A, 2016. Survival analysis for high-dimensional, heterogeneous medical data: Exploring feature extraction as an alternative to feature selection. Artif. Intell. Med. 72, 1–11. https://doi.org/10.1016/j.artmed.2016.07.004
Sekercioglu N, Fu R, Kim SJ, Mitsakakis N, 2021. Machine learning for predicting long-term kidney allograft survival: a scoping review. Ir. J. Med. Sci. https://doi.org/10.1007/s11845-020-02332-1
Spooner A, Chen E, Sowmya A, Sachdev P, Kochan NA, Trollor J, Brodaty H, 2020. A comparison of machine learning methods for survival analysis of high-dimensional clinical data for dementia prediction. Sci. Rep. 10, 20410. https://doi.org/10.1038/s41598-020-77220-w
Sraitih M, Jabrane Y, El Hassani AH, 2021. An automated system for ECG arrhythmia detection using machine learning techniques. J. Clin. Med. 10, 5450. https://doi.org/10.3390/jcm10225450
Stephane Fotso, 2019. PySurvival: Open source package for Survival Analysis modeling. https://www.pysurvival.io.
Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW, 2010. Assessing the performance of prediction models: A framework for traditional and novel measures. Epidemiology. https://doi.org/10.1097/EDE.0b013e3181c30fb2
Uno H, Cai T, Pencina MJ, D’Agostino RB, Wei LJ, 2011. On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data. Stat. Med. 30, 1105–1117. https://doi.org/10.1002/sim.4154
Van Rossum G, Drake FL, 2009. Python 3 Reference Manual. Scotts Val. CA Creat.
Wright MN, Dankowski T, Ziegler A, 2017. Unbiased split variable selection for random survival forests using maximally selected rank statistics. Stat. Med. 36, 1272–1284. https://doi.org/10.1002/sim.7212
Yu CN, Greiner R, Lin HC, Baracos V, 2011. Learning patient-specific cancer survival distributions as a sequence of dependent regressors, in: Advances in Neural Information Processing Systems 24: 25th Annual Conference on Neural Information Processing Systems 2011, NIPS 2011.
Zhu W, Xie L, Han J, Guo X, 2020. The application of deep learning in cancer prognosis prediction. Cancers (Basel). https://doi.org/10.3390/cancers12030603

Table 1

Overview of independent and outcome variables.

Biometric variables

height, weight, body mass index (BMI), age, sex.

Preoperative variables

past medical history (cardiovascular, pulmonary, metabolic and renal preconditions), tumor diagnosis, preceding malignant disease, cTNM classification, histology (Laurén type, signet cell component), neoadjuvant therapy (components, radiotherapy, completeness).

(Intra-)operative variables

time between diagnosis and resection, type of operation, extent of resection, anatomical reconstruction, duration of surgery, intraoperative complication, blood loss and transfusion.

Postoperative variables

days on ICU and ward, postoperative complications (according to Clavien-Dindo and additional 38 binarily classified types), pTNM classification, lymph node ratio (positive lymph nodes divided by resected), grading, R-status, histology (Laurén type, signet cell component, tumor regression), post-discharge problems.

Outcome variables

vital status, overall survival, no evidence of disease, time until tumor relapse, 30-day mortality and in-hospital mortality.

Table 2

Overview of selected patient characteristics and imputation methods. ASA = American Society of Anesthesiology, ICU = intensive care unit.

n = 1,360	Mean / Median / Frequency, (95% confidence interval)	Missingness
Biometric variables
Weight	77.7 kg (54.0 - 104.0)	0.0%
Height	1.72 m (1.58 - 1.87)	1.0%
Age	62.35 years (41.0 - 80.0)	0.0%
Sex	male (71.4%), female (28.6%)	0.0%
Preoperative variables
ASA score	1 (1.7%), 2 (47.9%), 3 (46.9%), 4 (2.3%)	1.2%
Past medical history	see Supplementary Table 1	max. 0.2%
cT status	1 (7.1%), 2 (20.1%), 3 (57.2%), 4 (10.9%)	4.7%
cN status	0 (35.7%), 1 (61.5%), + (0.4%)	2.4%
cM status	0 (87.7%), 1 (11.9%)	0.4%
Neoadjuvant therapy	yes (55.0%), no (45.0%)	0.0%
- Chemotherapeutics	see Supplementary Table 1	max. 2.1%
- Radiation	yes (3.5%), no (96.2%)	0.4%
(Intra-)operative variables
Time diagnosis to resection	82.84 days (11.0 - 168.1)	0.1%
One vs. two cavity surgery	one (71.1%), two (28.9%)	0.0%
Operation type	Subtotal gastrectomy (21.5%) Total gastrectomy (26.4%) Transhiatal extended gastrectomy (21.8%) Ivor-Lewis esophagectomy (27.8%) Other types (2.6%)	0.0%
(Locally) extended resection	yes (28.5%), no (71.2%)	0.3%
Intraoperative complications	yes (7.8%), no (92.1%)	0.1%
Duration of surgery	275.29 minutes (150.0 - 479.3)	0.9%
Intraoperative blood loss	622.93 milliliters (100.0 - 1,500.0)	19.2%
Postoperative variables
Duration of stay on ICU	6.89 days (0.0 - 29.0)	4.7%
Duration of stay on ward	20.26 days (9.0 - 49.6)	0.7%
Postoperative complications	see Supplementary Table 1	max. 0.2%
pT status	0/1 (22.3%), 2 (14.2%), 3 (47.3%), 4 (16.2%)	0.0%
pN status	0 (42.1%), 1 (16.3%), 2 (15.0%), 3 (26.5%)	0.1%
Lymph node ratio	18.25% (0.0 - 74.1)	0.1%
pM status	0 (89.7%), 1 (10.3%)	0.1%
R status	0 (81.5%), any 1/2/X (18.5%)	0.0%
Outcome variables
Status dead vs. alive/censored	alive (51.5%), dead (46.1%)	2.4%
Documented survival time	39.69 months (2.89 - 109.88)	1.0%

Table 3

Most important 20 predictors based on permutation-based importance scoring for random survival forest. ICU = intensive care unit, ASA = American Society of Anesthesiology.

Predictive Feature	Weight	Lower 95% CI	Upper 95% CI
Lymph node ratio (positive / total)	0.1183	0.1016	0.1350
Intraoperative blood loss	0.0060	-0.0009	0.0129
(y)pT4 status	0.0057	0.0002	0.0112
Age (at diagnosis)	0.0049	0.0035	0.0063
(y)pT3 status	0.0042	-0.0004	0.0088
Intraoperative peritoneal carcinosis	0.0040	0.0033	0.0047
Postoperative sepsis	0.0038	-0.0003	0.0079
cM+ status	0.0036	-0.0009	0.0081
Duration of in-hospital stay	0.0033	0.0028	0.0038
Duration of stay on ICU	0.0033	0.0004	0.0062
Major postoperative complications	0.0032	0.0020	0.0044
Any R+ status	0.0026	-0.0024	0.0076
Local R1 status	0.0021	0.0004	0.0038
(y)pM1 status	0.0020	0.0007	0.0033
(y)pT1 status	0.0014	0.0011	0.0017
Intraoperative blood transfusion	0.0013	0.0007	0.0019
Severe pre-existing diseases	0.0006	0.0002	0.0010
Siewert type I junction cancer	0.0006	-0.0004	0.0016
Cardiac postoperative adverse events	0.0005	0.0001	0.0009
ASA grade 3	0.0005	0.0000	0.0010

Machine Learning for Optimal Individual Survival Prediction in Resectable Upper Gastrointestinal Cancer

Status:

Version 1

Abstract

Figures

Introduction

Material And Methods