Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

oncology Oncology

Deep learning for automatic head and neck lymph node level delineation provides expert-level accuracy.

In Frontiers in oncology

BACKGROUND : Deep learning-based head and neck lymph node level (HN_LNL) autodelineation is of high relevance to radiotherapy research and clinical treatment planning but still underinvestigated in academic literature. In particular, there is no publicly available open-source solution for large-scale autosegmentation of HN_LNL in the research setting.

METHODS : An expert-delineated cohort of 35 planning CTs was used for training of an nnU-net 3D-fullres/2D-ensemble model for autosegmentation of 20 different HN_LNL. A second cohort acquired at the same institution later in time served as the test set (n = 20). In a completely blinded evaluation, 3 clinical experts rated the quality of deep learning autosegmentations in a head-to-head comparison with expert-created contours. For a subgroup of 10 cases, intraobserver variability was compared to the average deep learning autosegmentation accuracy on the original and recontoured set of expert segmentations. A postprocessing step to adjust craniocaudal boundaries of level autosegmentations to the CT slice plane was introduced and the effect of autocontour consistency with CT slice plane orientation on geometric accuracy and expert rating was investigated.

RESULTS : Blinded expert ratings for deep learning segmentations and expert-created contours were not significantly different. Deep learning segmentations with slice plane adjustment were rated numerically higher (mean, 81.0 vs. 79.6, p = 0.185) and deep learning segmentations without slice plane adjustment were rated numerically lower (77.2 vs. 79.6, p = 0.167) than manually drawn contours. In a head-to-head comparison, deep learning segmentations with CT slice plane adjustment were rated significantly better than deep learning contours without slice plane adjustment (81.0 vs. 77.2, p = 0.004). Geometric accuracy of deep learning segmentations was not different from intraobserver variability (mean Dice per level, 0.76 vs. 0.77, p = 0.307). Clinical significance of contour consistency with CT slice plane orientation was not represented by geometric accuracy metrics (volumetric Dice, 0.78 vs. 0.78, p = 0.703).

CONCLUSIONS : We show that a nnU-net 3D-fullres/2D-ensemble model can be used for highly accurate autodelineation of HN_LNL using only a limited training dataset that is ideally suited for large-scale standardized autodelineation of HN_LNL in the research setting. Geometric accuracy metrics are only an imperfect surrogate for blinded expert rating.

Weissmann Thomas, Huang Yixing, Fischer Stefan, Roesch Johannes, Mansoorian Sina, Ayala Gaona Horacio, Gostian Antoniu-Oreste, Hecht Markus, Lettmaier Sebastian, Deloch Lisa, Frey Benjamin, Gaipl Udo S, Distel Luitpold Valentin, Maier Andreas, Iro Heinrich, Semrau Sabine, Bert Christoph, Fietkau Rainer, Putz Florian

2023

artificial intelligence, autosegmentation, deep learning, head and neck, lymph node level, neural network, radiotherapy, target volume

Public Health Public Health

Applications of different machine learning approaches in prediction of breast cancer diagnosis delay.

In Frontiers in oncology

BACKGROUND : The increasing rate of breast cancer (BC) incidence and mortality in Iran has turned this disease into a challenge. A delay in diagnosis leads to more advanced stages of BC and a lower chance of survival, which makes this cancer even more fatal.

OBJECTIVES : The present study was aimed at identifying the predicting factors for delayed BC diagnosis in women in Iran.

METHODS : In this study, four machine learning methods, including extreme gradient boosting (XGBoost), random forest (RF), neural networks (NNs), and logistic regression (LR), were applied to analyze the data of 630 women with confirmed BC. Also, different statistical methods, including chi-square, p-value, sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC), were utilized in different steps of the survey.

RESULTS : Thirty percent of patients had a delayed BC diagnosis. Of all the patients with delayed diagnoses, 88.5% were married, 72.1% had an urban residency, and 84.8% had health insurance. The top three important factors in the RF model were urban residency (12.04), breast disease history (11.58), and other comorbidities (10.72). In the XGBoost, urban residency (17.54), having other comorbidities (17.14), and age at first childbirth (>30) (13.13) were the top factors; in the LR model, having other comorbidities (49.41), older age at first childbirth (82.57), and being nulliparous (44.19) were the top factors. Finally, in the NN, it was found that being married (50.05), having a marriage age above 30 (18.03), and having other breast disease history (15.83) were the main predicting factors for a delayed BC diagnosis.

CONCLUSION : Machine learning techniques suggest that women with an urban residency who got married or had their first child at an age older than 30 and those without children are at a higher risk of diagnosis delay. It is necessary to educate them about BC risk factors, symptoms, and self-breast examination to shorten the delay in diagnosis.

Dehdar Samira, Salimifard Khodakaram, Mohammadi Reza, Marzban Maryam, Saadatmand Sara, Fararouei Mohammad, Dianati-Nasab Mostafa

2023

breast cancer (BC), delay, extreme gradient boosting, logistic regression, machine learning, neural networks (NN), random forest (RF)

General General

Molecular features and predictive models identify the most lethal subtype and a therapeutic target for osteosarcoma.

In Frontiers in oncology

BACKGROUND : Osteosarcoma is the most common primary malignant bone tumor. The existing treatment regimens remained essentially unchanged over the past 30 years; hence the prognosis has plateaued at a poor level. Precise and personalized therapy is yet to be exploited.

METHODS : One discovery cohort (n=98) and two validation cohorts (n=53 & n=48) were collected from public data sources. We performed a non-negative matrix factorization (NMF) method on the discovery cohort to stratify osteosarcoma. Survival analysis and transcriptomic profiling characterized each subtype. Then, a drug target was screened based on subtypes' features and hazard ratios. We also used specific siRNAs and added a cholesterol pathway inhibitor to osteosarcoma cell lines (U2OS and Saos-2) to verify the target. Moreover, PermFIT and ProMS, two support vector machine (SVM) tools, and the least absolute shrinkage and selection operator (LASSO) method, were employed to establish predictive models.

RESULTS : We herein divided osteosarcoma patients into four subtypes (S-I ~ S-IV). Patients of S- I were found probable to live longer. S-II was characterized by the highest immune infiltration. Cancer cells proliferated most in S-III. Notably, S-IV held the most unfavorable outcome and active cholesterol metabolism. SQLE, a rate-limiting enzyme for cholesterol biosynthesis, was identified as a potential drug target for S-IV patients. This finding was further validated in two external independent osteosarcoma cohorts. The function of SQLE to promote proliferation and migration was confirmed by cell phenotypic assays after the specific gene knockdown or addition of terbinafine, an inhibitor of SQLE. We further employed two machine learning tools based on SVM algorithms to develop a subtype diagnostic model and used the LASSO method to establish a 4-gene model for predicting prognosis. These two models were also verified in a validation cohort.

CONCLUSION : The molecular classification enhanced our understanding of osteosarcoma; the novel predicting models served as robust prognostic biomarkers; the therapeutic target SQLE opened a new way for treatment. Our results served as valuable hints for future biological studies and clinical trials of osteosarcoma.

Zheng Kun, Hou Yushan, Zhang Yiming, Wang Fei, Sun Aihua, Yang Dong

2023

SQLE, cholesterol metabolism, drug target, molecular classification, osteosarcoma, predictive model

oncology Oncology

Quantitative analysis of artificial intelligence on liver cancer: A bibliometric analysis.

In Frontiers in oncology

OBJECTIVE : To provide the current research progress, hotspots, and emerging trends for AI in liver cancer, we have compiled a relative comprehensive and quantitative report on the research of liver disease using artificial intelligence by employing bibliometrics in this study.

METHODS : In this study, the Web of Science Core Collection (WoSCC) database was used to perform systematic searches using keywords and a manual screening strategy, VOSviewer was used to analyze the degree of cooperation between countries/regions and institutions, as well as the co-occurrence of cooperation between authors and cited authors. Citespace was applied to generate a dual map to analyze the relationship of citing journals and citied journals and conduct a strong citation bursts ranking analysis of references. Online SRplot was used for in-depth keyword analysis and Microsoft Excel 2019 was used to collect the targeted variables from retrieved articles.

RESULTS : 1724 papers were collected in this study, including 1547 original articles and 177 reviews. The study of AI in liver cancer mostly began from 2003 and has developed rapidly from 2017. China has the largest number of publications, and the United States has the highest H-index and total citation counts. The top three most productive institutions are the League of European Research Universities, Sun Yat Sen University, and Zhejiang University. Jasjit S. Suri and Frontiers in Oncology are the most published author and journal, respectively. Keyword analysis showed that in addition to the research on liver cancer, research on liver cirrhosis, fatty liver disease, and liver fibrosis were also common. Computed tomography was the most used diagnostic tool, followed by ultrasound and magnetic resonance imaging. The diagnosis and differential diagnosis of liver cancer are currently the most widely adopted research goals, and comprehensive analyses of multi-type data and postoperative analysis of patients with advanced liver cancer are rare. The use of convolutional neural networks is the main technical method used in studies of AI on liver cancer.

CONCLUSION : AI has undergone rapid development and has a wide application in the diagnosis and treatment of liver diseases, especially in China. Imaging is an indispensable tool in this filed. Mmulti-type data fusion analysis and development of multimodal treatment plans for liver cancer could become the major trend of future research in AI in liver cancer.

Xiong Ming, Xu Yaona, Zhao Yang, He Si, Zhu Qihan, Wu Yi, Hu Xiaofei, Liu Li

2023

Citespace, VOSviewer, artificial intelligence, bibliometrics, liver cancer

oncology Oncology

Artificial intelligence-based prediction of overall survival in metastatic renal cell carcinoma.

In Frontiers in oncology

BACKGROUND AND OBJECTIVES : Investigations of the prognosis are vital for better patient management and decision-making in patients with advanced metastatic renal cell carcinoma (mRCC). The purpose of this study is to evaluate the capacity of emerging Artificial Intelligence (AI) technologies to predict three- and five-year overall survival (OS) for mRCC patients starting their first-line of systemic treatment.

PATIENTS AND METHODS : The retrospective study included 322 Italian patients with mRCC who underwent systemic treatment between 2004 and 2019. Statistical analysis included the univariate and multivariate Cox proportional-hazard model and the Kaplan-Meier analysis for the prognostic factors' investigation. The patients were split into a training cohort to establish the predictive models and a hold-out cohort to validate the results. The models were evaluated by the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. We assessed the clinical benefit of the models using decision curve analysis (DCA). Then, the proposed AI models were compared with well-known pre-existing prognostic systems.

RESULTS : The median age of patients in the study was 56.7 years at RCC diagnosis and 78% of participants were male. The median survival time from the start of systemic treatment was 29.2 months; 95% of the patients died during the follow-up that finished by the end of 2019. The proposed predictive model, which was constructed as an ensemble of three individual predictive models, outperformed all well-known prognostic models to which it was compared. It also demonstrated better usability in supporting clinical decisions for 3- and 5-year OS. The model achieved (0.786 and 0.771) AUC and (0.675 and 0.558) specificity at sensitivity 0.90 for 3 and 5 years, respectively. We also applied explainability methods to identify the important clinical features that were found to be partially matched with the prognostic factors identified in the Kaplan-Meier and Cox analyses.

CONCLUSIONS : Our AI models provide best predictive accuracy and clinical net benefits over well-known prognostic models. As a result, they can potentially be used in clinical practice for providing better management for mRCC patients starting their first-line of systemic treatment. Larger studies would be needed to validate the developed model.

Barkan Ella, Porta Camillo, Rabinovici-Cohen Simona, Tibollo Valentina, Quaglini Silvana, Rizzo Mimma

2023

artificial intelligence, first-line treatment, machine learning, metastatic renal cell carcinoma, overall survival, predictive model

General General

Deep learning for detecting and elucidating human T-cell leukemia virus type 1 integration in the human genome.

In Patterns (New York, N.Y.)

Human T-cell leukemia virus type 1 (HTLV-1), a retrovirus, is the causative agent for adult T cell leukemia/lymphoma and many other human diseases. Accurate and high throughput detection of HTLV-1 virus integration sites (VISs) across the host genomes plays a crucial role in the prevention and treatment of HTLV-1-associated diseases. Here, we developed DeepHTLV, the first deep learning framework for VIS prediction de novo from genome sequence, motif discovery, and cis-regulatory factor identification. We demonstrated the high accuracy of DeepHTLV with more efficient and interpretive feature representations. Decoding the informative features captured by DeepHTLV resulted in eight representative clusters with consensus motifs for potential HTLV-1 integration. Furthermore, DeepHTLV revealed interesting cis-regulatory elements in regulation of VISs that have significant association with the detected motifs. Literature evidence demonstrated nearly half (34) of the predicted transcription factors enriched with VISs were involved in HTLV-1-associated diseases. DeepHTLV is freely available at https://github.com/bsml320/DeepHTLV.

Xu Haodong, Jia Johnathan, Jeong Hyun-Hwan, Zhao Zhongming

2023-Feb-10

HTLV-1, T cell leukemia/lymphoma, deep learning, motif, viral integration sites