Abstract
Early diagnosis of prostate cancer, the most common malignancy in men, can improve patient outcomes. Since the tissue sampling procedures are invasive and sometimes inconclusive, an alternative image-based method can prevent possible complications and facilitate treatment management. We aim to propose a machine-learning model for tumor grade estimation based on 68 Ga-PSMA-11 PET/CT images in prostate cancer patients. This study included 90 eligible participants out of 244 biopsy-proven prostate cancer patients who underwent staging 68Ga-PSMA-11 PET/CT imaging. The patients were divided into high and low-intermediate groups based on their Gleason scores. The PET-only images were manually segmented, both lesion-based and whole prostate, by two experienced nuclear medicine physicians. Four feature selection algorithms and five classifiers were applied to Combat-harmonized and non-harmonized datasets. To evaluate the model's generalizability across different institutions, we performed leave-one-center-out cross-validation (LOOCV). The metrics derived from the receiver operating characteristic curve were used to assess model performance. In the whole prostate segmentation, combining the ANOVA algorithm as the feature selector with Random Forest (RF) and Extra Trees (ET) classifiers resulted in the highest performance among the models, with an AUC of 0.78 and 083, respectively. In the lesion-based segmentation, the highest AUC was achieved by MRMR feature selector + Linear Discriminant Analysis (LDA) and Logistic Regression (LR) classifiers (0.76 and 0.79, respectively). The LOOCV results revealed that both the RF_ANOVA and ET_ANOVA models showed high levels of accuracy and generalizability across different centers, with an average AUC value of 0.87 for the ET_ANOVA combination. Machine learning-based analysis of radiomics features extracted from 68Ga-PSMA-11 PET/CT scans can accurately classify prostate tumors into low-risk and intermediate- to high-risk groups.
Similar content being viewed by others
Data availability
Not applicable.
Abbreviations
- ACC:
-
Accuracy
- ANOVA :
-
Analysis of Variance
- AUC:
-
Area Under the Curve
- CAD:
-
Computer-Aided Diagnosis
- ET:
-
Extra Trees
- FWHM:
-
Full Width at Half Maximum
- GLCM:
-
Gray-Level Co-Occurrence Matrix
- GLRLM:
-
Gray-Level Run Length Matrix
- GLZLM:
-
Gray-Level Zone-Length Matrix
- GS:
-
Gleason Score
- IBSI:
-
Image Biomarker Standardization Initiative
- KNN:
-
K-Nearest Neighbors
- KW:
-
Kruskal–Wallis test
- LDA:
-
Linear Discriminant Analysis
- LP:
-
Lesion of Prostate
- LR:
-
Logistic Regression
- ML :
-
Machine Learning
- MRMR:
-
Maximum Relevance Minimum Redundancy
- mpMRI:
-
Multiparametric Magnetic Resonance Imaging
- NGLDM:
-
Neighboring Gray-Level Dependence Matrix
- PCa:
-
Prostate Cancer
- PET/CT:
-
Positron Emission Tomography/Computed Tomography
- PR:
-
Precision
- PSA:
-
Prostate-Specific Antigen
- PSMA:
-
Prostate-Specific Membrane Antigen
- RF:
-
Random Forest
- ROC:
-
Receiver Operator Characteristic
- SUV:
-
Standardized Uptake Values
- TLG:
-
Total Lesion Glycolysis
- VOI:
-
Volume of Interest
- WP:
-
Whole Prostate
References
Plym A, Zhang Y, Stopsack KH, Delcoigne B, Wiklund F, Haiman C et al (2022) A healthy lifestyle in men at increased genetic risk for prostate cancer. Eur Urol 83:343
Chung LW, Isaacs WB, Simons JW (2007) Prostate cancer: biology, genetics, and the new therapeutics. Springer Science & Business Media, Totowa
Salam M. Principles and practice of Urology: JP Medical Ltd; 2013.
Eder M, Schäfer M, Bauder-Wüst U, Hull W-E, Wängler C, Mier W et al (2012) 68Ga-complex lipophilicity and the targeting property of a urea-based PSMA inhibitor for PET imaging. Bioconjug Chem 23(4):688–697
Beheshti M, Langsteger W, Rezaee A. PET/CT in cancer: an interdisciplinary approach to individualized imaging: Elsevier Health Sciences; 2017.
Emmett L, Willowson K, Violet J, Shin J, Blanksby A, Lee J (2017) Lutetium 177 PSMA radionuclide therapy for men with prostate cancer: a review of the current literature and discussion of practical aspects of therapy. Journal of medical radiation sciences 64(1):52–60
Donato P, Morton A, Yaxley J, Ranasinghe S, Teloken PE, Kyle S et al (2020) 68 Ga-PSMA PET/CT better characterises localised prostate cancer after MRI and transperineal prostate biopsy: Is 68 Ga-PSMA PET/CT guided biopsy the future? Eur J Nucl Med Mol Imaging 47:1843–1851
Schmuck S, Nordlohne S, von Klot C-A, Henkenberens C, Sohns JM, Christiansen H et al (2017) Comparison of standard and delayed imaging to improve the detection rate of [68 Ga] PSMA I&T PET/CT in patients with biochemical recurrence or prostate-specific antigen persistence after primary therapy for prostate cancer. Eur J Nucl Med Mol Imaging 44:960–968
Hatt M, Krizsan A, Rahmim A, Bradshaw T, Costa P, Forgacs A et al (2023) Joint EANM/SNMMI guideline on radiomics in nuclear medicine: Jointly supported by the EANM Physics Committee and the SNMMI Physics, Instrumentation and Data Sciences Council. Eur J Nucl Med Mol Imaging 50(2):352–375
Hajianfar G, Haddadi Avval A, Hosseini SA, Nazari M, Oveisi M, Shiri I, Zaidi H (2023) Time-to-event overall survival prediction in glioblastoma multiforme patients using magnetic resonance imaging radiomics. Radiol Med (Torino) 128:1521
Benoit-Cattin H. Texture analysis for magnetic resonance imaging: Texture Analysis Magn Resona; 2006.
Fusco R, Granata V, Grazzini G, Pradella S, Borgheresi A, Bruno A et al (2022) Radiomics in medical imaging: pitfalls and challenges in clinical management. Jpn J Radiol 40(9):919–929
Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, Van Stiphout RG, Granton P et al (2012) Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer 48(4):441–446
Hajianfar G, Kalayinia S, Hosseinzadeh M, Samanian S, Maleki M, Sossi V et al (2023) Prediction of Parkinson’s disease pathogenic variants using hybrid Machine learning systems and radiomic features. Physica Med 113:102647
Alongi P, Stefano A, Comelli A, Laudicella R, Scalisi S, Arnone G et al (2021) Radiomics analysis of 18F-Choline PET/CT in the prediction of disease outcome in high-risk prostate cancer: An explorative study on machine learning feature classification in 94 patients. Eur Radiol 31:4595–4605
Nai Y-H, Cheong DLH, Roy S, Kok T, Stephenson MC, Schaefferkoetter J et al (2023) Comparison of quantitative parameters and radiomic features as inputs into machine learning models to predict the Gleason score of prostate cancer lesions. Magn Reson Imaging 100:64–72
Zamboglou C, Carles M, Fechter T, Kiefer S, Reichel K, Fassbender TF et al (2019) Radiomic features from PSMA PET for non-invasive intraprostatic tumor discrimination and characterization in patients with intermediate-and high-risk prostate cancer-a comparison study with histology reference. Theranostics 9(9):2595
Zwanenburg A, Vallières M, Abdalah MA, Aerts HJ, Andrearczyk V, Apte A et al (2020) The image biomarker standardization initiative: standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 295(2):328–338
Nioche C, Orlhac F, Boughdad S, Reuzé S, Goya-Outi J, Robert C et al (2018) LIFEx: a freeware for radiomic feature calculation in multimodality imaging to accelerate advances in the characterization of tumor heterogeneity. Can Res 78(16):4786–4789
Shiri I, Amini M, Nazari M, Hajianfar G, Avval AH, Abdollahi H et al (2022) Impact of feature harmonization on radiogenomics analysis: Prediction of EGFR and KRAS mutations from non-small cell lung cancer PET/CT images. Comput Biol Med 142:105230
Hajianfar G, Avval AH, Sabouri M, Khateri M, Jenabi E, Geramifar P, et al., editors. ComBat Harmonization of Image Reconstruction Parameters to Improve the Repeatability of Radiomics Features. 2021 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC); 2021: IEEE.
Johnson WE, Li C, Rabinovic A (2007) Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8(1):118–127
Fortin J-P, Parker D, Tunç B, Watanabe T, Elliott MA, Ruparel K et al (2017) Harmonization of multi-site diffusion tensor imaging data. Neuroimage 161:149–170
Fortin J-P, Cullen N, Sheline YI, Taylor WD, Aselcioglu I, Cook PA et al (2018) Harmonization of cortical thickness measurements across scanners and sites. Neuroimage 167:104–120
Varghese B, Chen F, Hwang D, Palmer SL, De Castro Abreu AL, Ukimura O, et al., editors. Objective risk stratification of prostate cancer using machine learning and radiomics applied to multiparametric magnetic resonance images. Proceedings of the 11th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics; 2020.
Du D, Shiri I, Yousefirizi F, Salmanpour MR, Lv J, Wu H et al (2023) Impact of harmonization and oversampling methods on radiomics analysis of multi-center imbalanced datasets: Application to PET-based prediction of lung cancer subtypes. 71:209
DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44:837–845
Fendler WP, Eiber M, Beheshti M, Bomanji J, Calais J, Ceci F et al (2023) PSMA PET/CT: joint EANM procedure guideline/SNMMI procedure standard for prostate cancer imaging 2.0. Eur J Nucl Med Mol Imaging 50(5):1466–1486
Dehm SM, Tindall DJ (2020) Prostate Cancer: Cellular and Genetic Mechanisms of Disease Development and Progression. Springer Nature, Cham
Yu AC, Eng J (2020) One algorithm may not fit all: how selection bias affects machine learning performance. Radiographics 40(7):1932–1937
Shiri I, Salimi Y, Pakbin M, Hajianfar G, Avval AH, Sanaat A et al (2022) COVID-19 prognostic modeling using CT radiomic features and machine learning algorithms: Analysis of a multi-institutional dataset of 14,339 patients. Comput Biol Med 145:105467
Orlhac F, Eertink JJ, Cottereau A-S, Zijlstra JM, Thieblemont C, Meignan M et al (2022) A guide to ComBat harmonization of imaging biomarkers in multicenter studies. J Nucl Med 63(2):172–179
Nieri A, Manco L, Bauckneht M, Urso L, Caracciolo M, Donegani MI et al (2023) [18F] FDG PET-TC radiomics and machine learning in the evaluation of prostate incidental uptake. Expert Rev Med Devices 20(12):1183–1191
Yaxley JW, Raveenthiran S, Nouhaud F-X, Samartunga H, Yaxley AJ, Coughlin G et al (2019) Outcomes of primary lymph node staging of intermediate and high risk prostate cancer with 68Ga-PSMA positron emission tomography/computerized tomography compared to histological correlation of pelvic lymph node pathology. J Urol 201(4):815–820
Leung KH, Rowe SP, Leal JP, Ashrafinia S, Sadaghiani MS, Chung HW et al (2022) Deep learning and radiomics framework for PSMA-RADS classification of prostate cancer on PSMA PET. EJNMMI Res 12(1):1–15
Japkowicz N, Stephen S (2002) The class imbalance problem: A systematic study. Intelligent data analysis 6(5):429–449
Papp L, Spielvogel C, Grubmüller B, Grahovac M, Krajnc D, Ecsedi B et al (2021) Supervised machine learning enables non-invasive lesion characterization in primary prostate cancer with [68 Ga] Ga-PSMA-11 PET/MRI. Eur J Nucl Med Mol Imaging 48:1795–1805
Cysouw MC, Jansen BH, van de Brug T, Oprea-Lager DE, Pfaehler E, de Vries BM et al (2021) Machine learning-based analysis of [18 F] DCFPyL PET radiomics for risk stratification in primary prostate cancer. Eur J Nucl Med Mol Imaging 48:340–349
Mofrad FB, Zoroofi RA, Tehrani-Fard AA, Akhlaghpoor S, Sato Y (2014) Classification of normal and diseased liver shapes based on spherical harmonics coefficients. J Med Syst 38:1–9
Davnall F, Yip CS, Ljungqvist G, Selmi M, Ng F, Sanghera B et al (2012) Assessment of tumor heterogeneity: an emerging imaging tool for clinical practice? Insights Imaging 3:573–589
Chen X, Oshima K, Schott D, Wu H, Hall W, Song Y et al (2017) Assessment of treatment response during chemoradiation therapy for pancreatic cancer based on quantitative radiomic analysis of daily CTs: An exploratory study. PLoS ONE 12(6):e0178961
Qiu Q, Duan J, Yin Y (2020) Radiomics in radiotherapy: applications and future challenges. Precision Radiation Oncol 4(1):29–33
Dong X, Sun X, Sun L, Maxim PG, Xing L, Huang Y et al (2016) Early change in metabolic tumor heterogeneity during chemoradiotherapy and its prognostic value for patients with locally advanced non-small cell lung cancer. PLoS ONE 11(6):e0157836
Solari EL, Gafita A, Schachoff S, Bogdanović B, Villagrán Asiares A, Amiel T et al (2022) The added value of PSMA PET/MR radiomics for prostate cancer staging. Eur J Nucl Med Mol Imaging 49(2):527–538
Krajnc D, Spielvogel CP, Grahovac M, Ecsedi B, Rasul S, Poetsch N et al (2022) Automated data preparation for in vivo tumor characterization with machine learning. Front Oncol 12:1017911
Hambarde P, Talbar S, Mahajan A, Chavan S, Thakur M, Sable N (2020) Prostate lesion segmentation in MR images using radiomics based deeply supervised U-Net. Biocybernetics Biomed Eng 40(4):1421–1435
Gong L, Xu M, Fang M, He B, Li H, Fang X et al (2022) The potential of prostate gland radiomic features in identifying the Gleason score. Comput Biol Med 144:105318
Peng Y, Shen D, Liao S, Turkbey B, Rais-Bahrami S, Wood B et al (2015) MRI-based prostate volume-adjusted prostate-specific antigen in the diagnosis of prostate cancer. J Magn Reson Imaging 42(6):1733–1739
Karademir I, Shen D, Peng Y, Liao S, Jiang Y, Yousuf A et al (2013) Prostate volumes derived from MRI and volume-adjusted serum prostate-specific antigen: correlation with Gleason score of prostate cancer. AJR Am J Roentgenol 201(5):1041
Bai W, Fadil Y, Idrissi O, Dakir M, Debbagh A, Abouteib R (2021) The correlation between the gleason score of the biopsy and that of the prostatectomy patch. Annals Med Surg 63:102169
Shiri I, Amini M, Yousefirizi F, Vafaei Sadr A, Hajianfar G, Salimi Y et al (2023) Information fusion for fully automated segmentation of head and neck tumors from PET and CT images. Med Phys 51:319
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
The present study has been approved by Research Ethics Committees of Islamic Azad University - Science and Research Branch with the ethics code: IR.IAU.SRB.REC.1402.070.
Conflict of interest
The authors have no relevant financial or non-financial interests to disclose.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Khateri, M., Babapour Mofrad, F., Geramifar, P. et al. Machine learning-based analysis of 68Ga-PSMA-11 PET/CT images for estimation of prostate tumor grade. Phys Eng Sci Med (2024). https://doi.org/10.1007/s13246-024-01402-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13246-024-01402-3