Bias and Class Imbalance in Oncologic Data—Towards Inclusive and Transferrable AI in Large Scale Oncology Data Sets
Open accessThe DOI matches its Crossref record; 65 of the 66 references it cites were found in a registry or an index.
Abstract
Recent technological developments have led to an increase in the size and types of data in the medical field derived from multiple platforms such as proteomic, genomic, imaging, and clinical data. Many machine learning models have been developed to support precision/personalized medicine initiatives such as computer-aided detection, diagnosis, prognosis, and treatment planning by using large-scale medical data. Bias and class imbalance represent two of the most pressing challenges for machine learning-based problems, particularly in medical (e.g., oncologic) data sets, due to the limitations in patient numbers, cost, privacy, and security of data sharing, and the complexity of generated data. Depending on the data set and the research question, the methods applied to address class imbalance problems can provide more effective, successful, and meaningful results. This review discusses the essential strategies for addressing and mitigating the class imbalance problems for different medical data types in the oncologic domain.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures
![Figure 1 The number of publications on class imbalance listed in Pubmed by year for years 2012 to 2021 (blue = total publications per year on class imbalance, orange = total publications per year on class imbalance in oncology, grey = the number of publications on class imbalance pertaining to oncology as a percent of the total number of publications that year) [8].](https://figures.cochat.ai/paper-figures/ed9bef0892444e3bbe8f547cec448739/bias-and-class-imbalance-in-oncologic-datatowards-inclusive-and-transferrable-ai-figure-1.jpg)



Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Medical imaging such as chest X-ray (CXR), magnetic resonance imaging (MRI), ultrasound scan (USS), and computed tomography (CT) are exploited as input in ID-ai models [81,116,148].”
From Artificial intelligence bias in medical system designs: a systematic review (2023) · cites this paper as [148]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Regression shrinkage and selection via the lasso
Tibshirani, R. (1996) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1996.tb02080.x · Cited by 51,482 · cited 3×
“Tibshirani et al. [62] LASSO (Least Absolute Shrinkage and et al. [63] Graph-and kernel-based—omics data integration Hilario et al. [61] Class imbalance in proteomic biomarker studies overview Tibshirani et al. [62] LASSO (Least Absolute Shrinkage and Selection Operator) overview Yan et al. [63]…”
Focal Loss for Dense Object Detection
Lin, T.-Y., Goyal, P., Girshick, R., He, K., & Dollar, P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.324 · Cited by 30,581 · cited 1×
“In addition to the weighted cross-entropy loss and multiclass dice loss function, Lin et al. introduced the focal loss method to add a modulating factor (1-p i ) γ to the cross-entropy loss with tunable focusing parameter γ ≥ 0 [47].”
An introduction to ROC analysis
Fawcett, T. (2006) · Pattern Recognition Letters · doi:10.1016/j.patrec.2005.10.010 · Cited by 21,879
Dermatologist-level classification of skin cancer with deep neural networks
Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017) · Nature · doi:10.1038/nature21056 · Cited by 14,362 · cited 1×
“However, most published studies report positive findings and far less report on unsuccessful AI as exemplified in the context of, for example, skin cancer classification [30,31], sepsis prediction [32], the use of race correction in clinical algorithms [33], and notably the COVID-19 pandemic [34].”
BackgroundV-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation
Milletari, F., Navab, N., & Ahmadi, S.-A. (2016) · 2016 Fourth International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2016.79 · Cited by 11,931 · cited 1×
“Another class loss function to handle class imbalance in data for the medical image segmentation is the multiclass dice loss function [46].”
An introduction to variable and feature selection
GuyonIsabelle, & ElisseeffAndré (2003) · Journal of Machine Learning Research · doi:10.5555/944919.944968 · Cited by 7,715
An Introduction to Variable and Feature Selection.
Guyon, I., & Elisseeff, A. (2003) · Cited by 6,576
Sagi, O., & Rokach, L. (2018) · WIREs Data Mining and Knowledge Discovery · doi:10.1002/widm.1249 · Cited by 3,103 · cited 2×
“Ensemble learning combines multiple weak learners to acquire a robust model with better performance by using various mechanisms such as voting and boosting [51,52].”
BackgroundDong, X., Yu, Z., Cao, W., Shi, Y., & Ma, Q. (2020) · Frontiers of Computer Science · doi:10.1007/s11704-019-8208-z · Cited by 1,928 · cited 1×
“Ensemble learning combines multiple weak learners to acquire a robust model with better performance by using various mechanisms such as voting and boosting [51,52].”
BackgroundHidden in Plain Sight — Reconsidering the Use of Race Correction in Clinical Algorithms
Vyas, D. A., Eisenstein, L. G., & Jones, D. S. (2020) · New England Journal of Medicine · doi:10.1056/nejmms2004740 · Cited by 1,731 · cited 1×
“However, most published studies report positive findings and far less report on unsuccessful AI as exemplified in the context of, for example, skin cancer classification [30,31], sepsis prediction [32], the use of race correction in clinical algorithms [33], and notably the COVID-19 pandemic [34].”
BackgroundUsing Random Forest to Learn Imbalanced Data
Chen, C., Liaw, A. I., & Breiman, L. (2004) · UC Berkeley · Cited by 1,340
Bias in data‐driven artificial intelligence systems—An introductory survey
Ntoutsi, E., Fafalios, P., Gadiraju, U., Iosifidis, V., Nejdl, W., Vidal, M., Ruggieri, S., Turini, F., Papadopoulos, S., Krasanakis, E., et al. (2020) · WIREs Data Mining and Knowledge Discovery · doi:10.1002/widm.1356 · Cited by 1,043 · cited 1×
“The algorithmic bias here is defined as the prejudice embedded in the decision resulting from an AI system in favor or against a person, group, or thing that is taken into account to be unfair [14].”
BackgroundWong, A., Otles, E., Donnelly, J. P., Krumm, A., McCullough, J., DeTroyer-Cooley, O., Pestrue, J., Phillips, M., Konye, J., Penoza, C., et al. (2021) · JAMA Internal Medicine · doi:10.1001/jamainternmed.2021.2626 · Cited by 995 · cited 1×
“However, most published studies report positive findings and far less report on unsuccessful AI as exemplified in the context of, for example, skin cancer classification [30,31], sepsis prediction [32], the use of race correction in clinical algorithms [33], and notably the COVID-19 pandemic [34].”
A survey on addressing high-class imbalance in big data
Leevy, J. L., Khoshgoftaar, T. M., Bauder, R. A., & Seliya, N. (2018) · Journal of Big Data · doi:10.1186/s40537-018-0151-6 · Cited by 759
Larrazabal, A. J., Nieto, N., Peterson, V., Milone, D. H., & Ferrante, E. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1919012117 · Cited by 685 · cited 3×
“…most significant issue inhibiting the harnessing of data for meaningful and reproducible limits data , privacy in significant image class machine minority artificial whose predicted class probabilities are geared toward the majority class ignoring the significance of minority classes [12,13].”
A deep learning-based radiomics model for prediction of survival in glioblastoma multiforme
Lao, J., Chen, Y., Li, Z.-C., Li, Q., Zhang, J., Liu, J., & Zhai, G. (2017) · Scientific Reports · doi:10.1038/s41598-017-10649-8 · Cited by 630 · cited 2×
“Lao et al. proposed a deep learning-based radiomics model for survival prediction of Glioblastoma Multiforme (GBM) [66], wherein deep and handcrafted features were extracted from multi-modality MR images employing the LASSO regression model for the feature selection.”
Brain Tumor Segmentation and Radiomics Survival Prediction: Contribution to the BRATS 2017 Challenge
Isensee, F., Kickingereder, P., Wick, W., Bendszus, M., & Maier-Hein, K. H. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-75238-9_25 · Cited by 527 · cited 3×
“[40] Deep learning, nnU-net Brain tumor segmentation”
BackgroundInfluentialDaneshjou, R., Smith, M. P., Sun, M. D., Rotemberg, V., & Zou, J. (2021) · JAMA Dermatology · doi:10.1001/jamadermatol.2021.3129 · Cited by 403
Machine learning and integrative analysis of biomedical big data
Mirza, B., Wang, W., Wang, J., Choi, H., Chung, N. C., & Ping, P. (2019) · Genes · doi:10.3390/genes10020087 · Cited by 380 · cited 5×
“…the precision-recall curve (auPRC), Matthews correlation coefficient (MCC), F-score, and the geometric mean (Gmean) of sensitivity and specificity [59,63] however may not yield adequate re-sults. imbalance-specific evaluation measures such as the area under the precision-recall , Matthews…”
InfluentialA Comparative Performance Analysis of Data Resampling Methods on Imbalance Medical Data
Khushi, M., Shaukat, K., Alam, T. M., Hameed, I. A., Uddin, S., Luo, S., Yang, X., & Reyes, M. C. (2021) · IEEE Access · doi:10.1109/access.2021.3102399 · Cited by 335 · cited 8×
“Furthermore, the machine learning models treat minority classes as noise; therefore, noise may be incorrectly classified [15].”
InfluentialBelenguer, L. (2022) · AI and Ethics · doi:10.1007/s43681-022-00138-8 · Cited by 289
Landers, R. N., & Behrend, T. S. (2023) · American Psychologist · doi:10.1037/amp0000972 · Cited by 233 · cited 2×
“The number and importance of publications about the It impacts the transfer of findings between the lab and the clinic and limits improvement in patient outcomes [9]. conclusions.”
Sex-Specific Differences in Glioblastoma
Carrano, A., Juarez, J. J., Incontri, D., Ibarra, A., & Guerrero Cazares, H. (2021) · Cells · doi:10.3390/cells10071783 · Cited by 185 · cited 1×
“Oncologic disciplines where this has been highlighted include large patient volume oncologic sites such as lung cancer [20], breast cancer [18], and head and neck cancer [21] but also lower volume sites such as central nervous system malignancies [22,23].”
BackgroundDROCC: Deep Robust One-Class ClassificationPreprint
Goyal, S., Raghunathan, A., Jain, M., Simhadri, H. V., & Jain, P. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.12718 · Cited by 180 · cited 1×
“Various models, such as support vector machine (SVM) and isolation forest, can be built by one class learning, while some models, such as decision trees and Naïve Bayes, cannot [41].”
Learning Pattern Classification Tasks with Imbalanced Data Sets
Hoang, G., Bouzerdoum, A., & Lam, S. (2009) · Pattern Recognition · doi:10.5772/7544 · Cited by 178 · cited 4×
“However, obtaining balanced data in medical data sets of all subtypes is challenging secondary to limitations in patient numbers, cost, privacy a d security of data sharing, and the complexity of generat d data, resulting in significant efforts to achi ve representative data sets [10,11].”
BackgroundInfluentialApproaches to dimensionality reduction in proteomic biomarker studies
Hilario, M., & Kalousis, A. (2007) · Briefings in Bioinformatics · doi:10.1093/bib/bbn005 · Cited by 165
Handling imbalanced medical image data: A deep-learning-based one-class classification approach
Gao, L., Zhang, L., Liu, C., & Wu, S. (2020) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2020.101935 · Cited by 162 · cited 1×
“Gao et al. proposed a deep-learning-based one-class classification approach to handle imbalanced class data for medical images such as breast MRIs and space-occupying kidney lesion (SOKL) data sets [42].”
Megahed, F. M., Chen, Y., Megahed, A., Ong, Y. J., Altman, N., & Krzywinski, M. I. (2021) · Nature Methods · doi:10.1038/s41592-021-01302-4 · Cited by 139 · cited 2×
“…underrepresented. distributions within data sets result in Class imbalance in clinical, imaging, and omics data sets is the rule, not the exception [5]. majority class, or the subset of the data overrepresented, and a minority class In the context of clinical data, this may be the result of the…”
The evolving landscape of sex-based differences in lung cancer: a distinct disease in women
Ragavan, M. V., & Patel, M. I. (2022) · European Respiratory Review · doi:10.1183/16000617.0100-2021 · Cited by 101 · cited 2×
“They have also been shown to affect data collected at all levels, clinical [6], imaging [13], and biospecimen [20], and in turn, analysis of downstream genomic and proteomic data sets (Figure 2) [8].”
BackgroundRajaraman, S., Ganesan, P., & Antani, S. (2022) · PLOS ONE · doi:10.1371/journal.pone.0262838 · Cited by 86 · cited 1×
“Machine learning models/ artificial intelligence (AI) systems considering all data, therefore, produce biased classifiers whose predicted class probabilities are geared toward the majority class ignoring the significance of minority classes [12,13].”
BackgroundWu, A., Li, Y., Qi, M., Lu, X., Jia, Q., Guo, F., Dai, Z., Liu, Y., Chen, C., Zhou, L., et al. (2020) · Oral Oncology · doi:10.1016/j.oraloncology.2020.104625 · Cited by 79 · cited 1×
“They constructed the RadModel and RadDosModel for this purpose [67], extracted shape-based, statistical, texture-based radiomic features using Pyradiomics tool, and subsequently selected relevant and essential features which were then compressed using principal component analysis technique (PCA).”
MethodGender and race interact to influence survival disparities in head and neck cancer
Mazul, A. L., Naik, A. N., Zhan, K. Y., Stepan, K. O., Old, M. O., Kang, S. Y., Nakken, E. R., & Puram, S. V. (2020) · Oral Oncology · doi:10.1016/j.oraloncology.2020.105093 · Cited by 74 · cited 1×
“Oncologic disciplines where this has been highlighted include large patient volume oncologic sites such as lung cancer [20], breast cancer [18], and head and neck cancer [21] but also lower volume sites such as central nervous system malignancies [22,23].”
BackgroundCan we trust AI not to further embed racial bias and prejudice?
Noor, P. (2020) · BMJ · doi:10.1136/bmj.m363 · Cited by 71
Höhn, J., Hekler, A., Krieghoff-Henning, E., Kather, J. N., Utikal, J. S., Meier, F., Gellrich, F. F., Hauschild, A., French, L., Schlager, J. G., et al. (2021) · Journal of Medical Internet Research · doi:10.2196/20708 · Cited by 66
Rising to the challenge of bias in health care AI
Cho, M. K. (2021) · Nature Medicine · doi:10.1038/s41591-021-01577-2 · Cited by 55
Massey, S. C., Whitmire, P., Doyle, T. E., Ippolito, J. E., Mrugała, M. M., Hu, L., Canoll, P., Anderson, A. R. A., Wilson, M. A., Fitzpatrick, S. M., et al. (2020) · Cancer Letters · doi:10.1016/j.canlet.2020.07.030 · Cited by 53 · cited 1×
“Oncologic disciplines where this has been highlighted include large patient volume oncologic sites such as lung cancer [20], breast cancer [18], and head and neck cancer [21] but also lower volume sites such as central nervous system malignancies [22,23].”
BackgroundDigital Health COVID-19 Impact Assessment: Lessons Learned and Compelling Needs
Microsoft, Lee, P., Abernethy, A., Verily, Shaywitz, D., HealthTech, A., Gundlapalli, A., Prevention, U. S. C. F. D. C. A., Weinstein, J., Healthcare, M., et al. (2022) · NAM Perspectives · doi:10.31478/202201c · Cited by 49
Social Determinants of Health and Disparities in Cancer Care for Black People in the United States
Tucker‐Seeley, R. D. (2021) · JCO Oncology Practice · doi:10.1200/op.21.00229 · Cited by 46 · cited 7×
“Social determinants of health consist of the geographical and socioeconomic conditions where people are born, live, and age, and the drivers of these conditions with cancer outcome disparities reflecting these aspects [24].”
InfluentialCancer Classification with a Cost-Sensitive Naive Bayes Stacking Ensemble
Xiong, Y., Ye, M., & Wu, C. (2021) · Computational and Mathematical Methods in Medicine · doi:10.1155/2021/5556992 · Cited by 45 · cited 1×
“In a study, the embedding cost-sensitive naive Bayes (NB) classifier was used as a meta-learner to handle class imbalance for cancer gene expression data [49].”
Classification of Kidney Cancer Data Using Cost-Sensitive Hybrid Deep Learning Approach
Shon, H. S., Batbaatar, E., Kim, K. O., Cha, E. J., & Kim, K.-A. (2020) · Symmetry · doi:10.3390/sym12010154 · Cited by 43
Comparison of Two Main Approaches for Handling Imbalanced Data in Churn Prediction Problem
Nguyen, N. N., & Duong, A. T. (2021) · Journal of Advances in Information Technology · doi:10.12720/jait.12.1.29-35 · Cited by 43 · cited 1×
“The weighted cross-entropy loss equation [45] is defined as Equation (1).”
Rossi, L. A., Melstrom, L. G., Fong, Y., & Sun, V. C.-Y. (2021) · Journal of Surgical Oncology · doi:10.1002/jso.26413 · Cited by 35 · cited 1×
“Computerized patient-reported outcome measures and patient-generated data are on the rise [29] in clinical settings, although such large-scale data are likely to intersect with other sources of data imbalance.”
BackgroundMarcelino, A. C., Gozzi, B., Cardoso-Filho, C., Machado, H. D. C., Zeferino, L. C., & Vale, D. B. (2021) · BMC Cancer · doi:10.1186/s12885-021-08735-2 · Cited by 35 · cited 2×
“Notably, gender [16,17] and race bias [18,19] with class imbalances are increasingly highlighted in the literature.”
BackgroundYan, K. K., Zhao, H., & Pang, H. (2017) · BMC Bioinformatics · doi:10.1186/s12859-017-1982-4 · Cited by 34 · cited 4×
“Tibshirani et al. [62] LASSO (Least Absolute Shrinkage and et al. [63] Graph-and kernel-based—omics data integration Hilario et al. [61] Class imbalance in proteomic biomarker studies overview Tibshirani et al. [62] LASSO (Least Absolute Shrinkage and Selection Operator) overview Yan et al. [63]…”
InfluentialExploring the Performance of Resampling Strategies for the Class Imbalance Problem
García, V., Sánchez, J. S., & Mollineda, R. A. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-13022-9_54 · Cited by 34 · cited 1×
“They aim to balance data sets by utilizing re-sampling methods such as increasing the sample size of the minority classes (over-sampling) or reducing the sample size of the majority class (under-sampling) [15,65].”
MethodLeech, M. M., Weiss, J. E., Markey, C., & Loehrer, A. P. (2022) · Annals of Surgical Oncology · doi:10.1245/s10434-021-11160-1 · Cited by 33 · cited 1×
“Both SDOH and disability have been particularly challenging to capture robustly in data sets and, as a result, are often not included in AI analysis [24,25].”
BackgroundA simple model for glioma grading based on texture analysis applied to conventional brain MRI
Suárez-García, J. G., Hernández-López, J. M., Moreno-Barbosa, E., & Alonso, B. D. C. (2020) · PLoS ONE · doi:10.1371/journal.pone.0228972 · Cited by 30 · cited 1×
“In another study, Suarez-Garcia et al. also developed a basic model for glioma grading based on texture analysis using the under-sampling method to handle class imbalance data [38].”
Morshed, R. A., Reihl, S. J., Molinaro, A. M., Kakaizada, S., Young, J. S., Schulte, J. D., Butowski, N. A., Taylor, J. W., Bush, N. A. O., Aghi, M. K., et al. (2020) · Journal of Neuro-Oncology · doi:10.1007/s11060-020-03503-x · Cited by 28 · cited 1×
“Notably, gender [16,17] and race bias [18,19] with class imbalances are increasingly highlighted in the literature.”
BackgroundÖzdemir, B. C., Oertelt‐Prigione, S., Adjei, A. A., Borchmann, S., Haanen, J. B. A. G., Letsch, A., Mir, O., Quaas, A., Verhoeven, R. H. A., & Wagner, D. (2022) · Annals of Oncology · doi:10.1016/j.annonc.2021.11.011 · Cited by 28 · cited 1×
“Notably, gender [16,17] and race bias [18,19] with class imbalances are increasingly highlighted in the literature.”
BackgroundLiu, R., Hall, L., Bowyer, K. W., Goldgof, D. B., Gatenby, R. A., & Ahmed, K. B. (2017) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc.2017.8122802 · Cited by 27 · cited 1×
“…Synthetic Minority Oversampling Technique (SMOTE) to overcome with class imbalance problem for Glioblastoma Multiforme (GBM) brain tumor images obtained from Moffitt Cancer Research Center and TCGA (The Cancer Genome Atlas Program) T1-Contrast data set (NIH’s Cancer Imaging Archive Dataset) [37].”
Whale optimized mixed kernel function of support vector machine for colorectal cancer diagnosis
Zhao, D., Liu, H., Zheng, Y., He, Y., Lu, D., & Lyu, C. (2019) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2019.103124 · Cited by 27 · cited 1×
“To increase the predictive performance of colorectal cancer with the imbalanced data, Zhao et al. employed the RUSBoost algorithm for the colorectal cancer diagnosis [57].”
Gender and Sex-Related Differences in Normal Tissue Effects Induced by Platinum Compounds
Marcu, L. G. (2022) · Pharmaceuticals · doi:10.3390/ph15020255 · Cited by 25 · cited 1×
“Notably, gender [16,17] and race bias [18,19] with class imbalances are increasingly highlighted in the literature.”
Welch, M. L., McIntosh, C., McNiven, A., Huang, S. H., Zhang, B., Wee, L., Traverso, A., O’Sullivan, B., Hoebers, F. J. P., Dekker, A. L. A. J., et al. (2020) · Physica Medica · doi:10.1016/j.ejmp.2020.01.027 · Cited by 24 · cited 1×
“For head and neck cancer radiation therapy outcome predictions, Welch et al. (2020) employed the isolation forest algorithm, an ensemble of isolation trees constructed to deal with the class imbalance and detect data anomalies [43].”
Tang, X., Cai, L., Meng, Y., Gu, C., Yang, J., & Yang, J. (2021) · IEEE Access · doi:10.1109/access.2021.3070428 · Cited by 23 · cited 4×
“constructed multi-omics data with the transcriptome and functional proteomics data and employed a bagging-based ensemble learning method on breast cancer to avoid over-fitting problems [53].”
BackgroundMethodInfluentialFeghali, K. A. A., Buszek, S. M., Elhalawani, H. M., Chevli, N., Allen, P. K., & Chung, C. D. (2020) · Neuro-Oncology Practice · doi:10.1093/nop/npaa064 · Cited by 21 · cited 1×
“Notably, demographics and management can be drastically different between trial populations and the real world [27,28], resulting in heterogeneous data mirrored in data sets employed for analysis, training, testing, and validation (Figure 2).”
BackgroundLarge-scale labeling and assessment of sex bias in publicly available expression data
Flynn, E., Chang, A., & Altman, R. B. (2021) · BMC Bioinformatics · doi:10.1186/s12859-021-04070-2 · Cited by 20
Wang, S., Dong, D., Zhang, W., Hu, H., Li, H., Zhu, Y., Zhou, J., Shan, X., & Tian, J. (2021) · Medical Physics · doi:10.1002/mp.15094 · Cited by 20 · cited 1×
“Meanwhile, Wang et al. introduced a multi-layer perceptron (MLP)-based ensemble learning architecture for the classification of advanced gastric cancer (AGC) [55].”
Jung, H., Lu, M., Quan, M. L., Cheung, W. Y., Kong, S., Lupichuk, S., Feng, Y., & Xu, Y. (2022) · BMC Cancer · doi:10.1186/s12885-022-09333-6 · Cited by 18 · cited 1×
“Notably, demographics and management can be drastically different between trial populations and the real world [27,28], resulting in heterogeneous data mirrored in data sets employed for analysis, training, testing, and validation (Figure 2).”
BackgroundBrown, W. E., Sung, K., Aleman, D. M., Moreno‐Centeno, E., Purdie, T. G., & McIntosh, C. (2018) · Medical Physics · doi:10.1002/mp.12757 · Cited by 17 · cited 2×
“Their proposed approach obtains strong performance results with high classification accuracy [36].”
BackgroundPredicting liver cancers using skewed epidemiological data
Li, J., Tao, Y., Cong, H., Zhu, E., & Cai, T. (2022) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2021.102234 · Cited by 17 · cited 1×
“Furthermore, to predict the occurrence of liver cancer with skewed epidemiological data and tackle the class imbalance problem, Li et al. introduced two different under-sampling methods based on K-Means++ and learning vector quantization (LVQ) [39].”
Prognostic prediction of histopathological images by local binary patterns and RUSBoost
Urdal, J., Engan, K., Kvikstad, V., & Janssen, E. A. M. (2017) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco.2017.8081630 · Cited by 16 · cited 1×
“A RUSBoost classifier was also used by Urdal et al. to solve and alleviate the class imbalance problems for the prognosis prediction of superficial urothelial carcinomas [58].”
Williamson, C. W., Nelson, T. J., Thompson, C. A., Vitzthum, L. K., Zakeri, K., Riviere, P. J., Bryant, A. K., Sharabi, A. B., Zou, J., & Mell, L. K. (2022) · Clinical Cancer Research · doi:10.1158/1078-0432.ccr-21-2468 · Cited by 11 · cited 1×
“This heterogeneity in “big data” is reflected in publicly available data sets in all types of medical settings and clinical contexts [6,7].”
BackgroundBose, P., Sleeman, W. C., Syed, K., Hagan, M., Palta, J. R., Kapoor, R., & Ghosh, P. (2021) · ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics (ACM-BCB) · doi:10.1145/3459930.3469518 · Cited by 7 · cited 1×
“…process for saving time and cost as well as overcoming errors in the radiation oncology incident learning system, Bose et al. applied convolutional neural networks (CNN) and long short-term memory (LSTM) with minority and random over-sampling strategies to handle high-class imbalanced data [35].”
A simple model for glioma grading based on texture analysis applied to conventional brain MRIPreprint
Suárez-García, J. G., Hernández-López, J. M., Moreno-Barbosa, E., et al. (2020) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2020.01.29.924712 · Cited by 6
Li, S., Natwick, T., Liu, J., Morrison, V. A., Vidito, S. I., Werther, W., Yusuf, A., & Usmani, S. Z. (2019) · Journal of Geriatric Oncology · doi:10.1016/j.jgo.2019.01.007 · Cited by 3 · cited 1×
“Class imbalance carries significant implications concerning these domains, and studies have looked at computational approaches, including natural language processing in unstructured data to expand data sets to capture disability as a feature [26].”
Research works printed but not linked: 4
- National Library of Medicine Available online: https://pubmed.ncbi.nlm.nih.gov (accessed on 1 December 2021)
- Abd Elrahman S.M. Abraham A. A Review of Class Imbalance Problem J. Netw. Innov. Comput. 2013 1 332 340
- Jaeger P.F. Kohl S.A.A. Bickelhaupt S. Isensee F. Kuder T.A. Schlemmer H.-P. Maier-Hein K.H. Retina U-Net: Embarrassingly Simple Exploitation of Segmentation Supervision for Medical Object Detection Proceedings of the Machine Learning for Health NeurIPS Workshop, Proceedings of Machine Learning Research 2020 171 183 Available online: http://proceedings.mlr.press/v116/jaeger20a/jaeger20a.pdf (accessed on 1 December 2021)
- Le D.N.T. Le H.X. Ngo L. Ngo H.T. Transfer learning with class-weighted and focal loss function for automatic skin cancer classification arXiv 2020 2009.05977
Cited by
Fan, Z., Yan, Z., & Wen, S. (2023) · Sustainability · doi:10.3390/su151813493 · Cited by 346
“This risk is not merely theoretical; recent studies have shown that biases in AI models can have significant real-world impacts, such as reinforcing existing social inequalities [108,109].”
BackgroundAI-driven biomarker discovery: enhancing precision in cancer diagnosis and prognosis
Alum, E. U. (2025) · Discover Oncology · doi:10.1007/s12672-025-02064-7 · Cited by 208
Challenges and limitations of synthetic minority oversampling techniques in machine learning
Alkhawaldeh, I. M., Albalkhi, I., & Naswhan, A. J. (2023) · World Journal of Methodology · doi:10.5662/wjm.v13.i5.373 · Cited by 179
A review of cancer data fusion methods based on deep learning
Zhao, Y., Li, X., Zhou, C., Peng, H., Zheng, Z., Chen, J., & Ding, W. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102361 · Cited by 95
Gürcan, F., & Soylu, A. (2024) · Cancers · doi:10.3390/cancers16193417 · Cited by 133
“All experiments were conducted using the Stratified K-Fold (5-Fold) Cross-Validation method [1,9,67].”
Method
Metadata
- Authors
- Tasci, E., Zhuge, Y., Camphausen, K., & Krauze, A. V.
- Affiliations
- Center for Cancer Research, National Cancer Institute, NIH, Building 10, Bethesda, MD 20892, USA
- National Institutes of Health (US)
- Corresponding author
- Krauze, Andra V.
- Venue
- Cancers
- Publisher
- MDPI AG
- Published
- 2022-06-12
- Volume
- 14
- Issue
- 12
- Pages
- 2897
- ISSN
- 2072-6694 (ISSN-L)
- PMC
- PMC9221277
- OpenAlex
- W4282548202
- Cited by
- 117
- References
- 66
- Citations by year
- 2026: 38 · 2025: 21 · 2024: 28 · 2023: 23 · 2022: 7
- FWCI
- 1.38
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Cancers (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Funders
- NCI NIH intramural program (ZID BC 010990)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Imbalanced Data Classification Techniques
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Keywords
- class imbalance, algorithmic bias, medical datasets, precision medicine, computer-aided diagnosis, transfer learning, data privacy, data sharing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.