Prediction of post-radiotherapy locoregional progression in HPV-associated oropharyngeal squamous cell carcinoma using machine-learning analysis of baseline PET/CT radiomics

Haider, Stefan P.; Sharaf, Kariem; Zeevi, Tal; Baumeister, Philipp; Reichel, Christoph; Forghani, Reza; Kann, Benjamin H.; Petukhova, Alexandra; Judson, Benjamin L.; Prasad, Manju L.; Liu, Chi; Burtness, Barbara; Mahajan, Amit; Payabvash, Seyedmehdi

doi:10.1016/j.tranon.2020.100906

Cited by 22 publications

(21 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, the limited size of the cohort limited meaningful subgroup analyses in this study and did not allow us to test this hypothesis. Other published models of LF or LRF in the head and neck region have tended to be more restrictive with the primary sites included, e.g., oropharyngeal (17)(18)(19), laryngeal/hypopharyngeal (20), hypopharyngeal (22). Still, our model compared favorably to previously published models predicting LF or LRF of HNSCC (as discussed in the introduction).…”

Section: Discussionmentioning

confidence: 48%

“…Still, our model compared favorably to previously published models predicting LF or LRF of HNSCC (as discussed in the introduction). Additionally, our analysis is limited by a lack of comparison of delta radiomic features to other known imaging modalities, such as PET/CT (18)(19)(20)(21), that have been shown to be informative for LF prediction of HNSCC. Therefore, we cannot comment on the superiority of this radiomic method in comparison to other radiomic methods that use features from other modalities.…”

Section: Discussionmentioning

confidence: 99%

“…This process of extracting and quantifying numerous features from medical images is termed radiomics (10,11). Prior studies have shown promising results with the correlation of radiomic features of baseline CT (12)(13)(14)(15)(16)(17), [18F] fluoro-2-deoxy-D-glucose (FDG) PET/CT (18)(19)(20)(21), and MRI images (22) with LF or loco-regional failure (LRF). Of those that were externally validated, Folkert et al showed that an FDG-PET/CT-based multiparameter logistic regression model achieved an AUC of 0.68 and sensitivity of 67% for the prediction of LF of oropharyngeal primaries (18), Bogowicz et al showed that a mixed model using CT-based radiomics of both the primary and lymph nodes achieved an AUC of 0.67 for predicting LRF (12), Giraud et al showed an XGboost model using CT-based radiomics of oropharyngeal primaries to achieve an AUC of 0.68 and sensitivity of 42% at predicting LRF (17), and Legers et al showed CT-based radiomic risk models of tumor rim sub-volumes to achieve an AUC of 0.63 to 0.65 at predicting LRF (15).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features

Morgan¹,

Wang²,

Dohopolski³

et al. 2021

Quant Imaging Med Surg

View full text Add to dashboard Cite

Background: Local failure (LF) following chemoradiation (CRT) for head and neck cancer is associated with poor overall survival. If machine learning techniques could stratify patients at risk of treatment failure based on baseline and intra-treatment imaging, such a model could facilitate response-adapted approaches to escalate, de-escalate, or switch therapy.Methods: A 1:2 retrospective case control cohort of patients treated at a single institution with definitive radiotherapy for head and neck cancer who failed locally, in-field at a primary or nodal structure were included. Radiomic features were extracted from baseline CT and CBCT scans at fractions 1 and 21 (delta) of radiotherapy with PyRadiomics and were selected for by: reproducibility (intra-class correlation coefficients ≥0.95), redundancy [maximum relevance and minimum redundancy (mRMR)], and informativeness [recursive feature elimination (RFE)]. Separate models predicting LF of primaries or nodes were created using the explainable boosting machine (EBM) classifier with 5-fold cross-validation for (I) clinical only, (II) radiomic only (CT1 and delta features), and (III) fused models (clinical + radiomic). Twenty-five iterations were performed, and predicted scores were averaged with a parallel ensemble design. Receiver operating characteristic curves were compared between models with paired-samples t-tests. Results:The fused ensemble model for primaries (using clinical, CT1, and delta features) achieved an AUC of 0.871 with a sensitivity of 78.3% and specificity of 90.9% at the maximum Youden J statistic.The fused ensemble model trended towards improvement when compared to the clinical only ensemble model (AUC =0.788, P=0.134) but reached significance when compared to the radiomic ensemble model (AUC =0.770, P=0.017). The fused ensemble model for nodes achieved an AUC of 0.910 with a sensitivity of 100.0% and specificity of 68.0%, which also trended towards improvement when compared to the clinical model (AUC =0.865, P=0.080). Conclusions:The fused ensemble EBM model achieved high discriminatory ability at predicting LF for head and neck cancer in independent primary and nodal structures. Although an additive benefit of delta radiomics over clinical factors could not be proven, the results trended towards improvement with the fused ensemble model, which are promising and worthy of prospective investigation in a larger cohort.

show abstract

Section: Discussionmentioning

confidence: 48%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features

Morgan¹,

Wang²,

Dohopolski³

et al. 2021

Quant Imaging Med Surg

View full text Add to dashboard Cite

show abstract

“…As survival rates in HNSCC patients have shown little improvement over the past decades, novel approaches in treatment development and precision diagnostics and prognostication are paramount for improving outcomes. Omics -research, based on quantitative analysis of large sets of biologic data, yielded promising results [ 28 , 29 , 30 ]. However, CSCs, which represent only a small population of HNSCC cells, may escape wholistic analysis approaches such as radiomics, genomics, or proteomics.…”

Section: Discussionmentioning

confidence: 99%

Discrimination of Cancer Stem Cell Markers ALDH1A1, BCL11B, BMI-1, and CD44 in Different Tissues of HNSCC Patients

et al. 2021

Self Cite

View full text Add to dashboard Cite

Cancer stem cells (CSCs) are accountable for the progress of head and neck squamous cell carcinoma (HNSCC). This exploratory study evaluated the expression of molecular CSC markers in different tissues of HNSCC patients. Tissue specimens of primary tumor, lymph node metastases and macroscopically healthy mucosa of 12 consecutive HNSCC patients, that were treated with surgery and adjuvant radio(chemo)therapy upon indication, were collected. Samples were assessed for the expression of p16 as a surrogate for HPV-related disease and different molecular stem cell markers (ALDH1A1, BCL11B, BMI-1, and CD44). In the cohort, seven patients had HPV-related HNSCC; six thereof were oropharyngeal squamous cell carcinoma. While expression of BMI-1 and BCL11B was significantly lower in healthy mucosa than both tumor and lymph node metastasis, there were no differences between tumor and lymph node metastasis. In the HPV-positive sub-cohort, these differences remained significant for BMI-1. However, no significant differences in these three tissues were found for ALDH1A1 and CD44. In conclusion, this exploratory study shows that CSC markers BMI-1 and BCL11B discriminate between healthy and cancerous tissue, whereas ALDH1A1 and CD44 were expressed to a comparable extent in healthy mucosa and cancerous tissues.

show abstract

“…Die Ergebnisse dieser Arbeiten deuten darauf hin, dass die zusätzliche Verwendung von KI der gegenwärtigen diagnostischen Aufarbeitung oro-, naso- und hypopharyngealer sowie oraler und laryngealer Karzinome hinsichtlich Prognostik bzw. Prädiktion der jeweiligen Studienendpunkte überlegen ist [ 11 , 12 , 18 ]. Im Speziellen belegen kürzlich veröffentlichte Untersuchungen der Autoren dieses Artikels einen additiven Nutzen im Hinzuziehen von PET-CT-Radiomics-Markern im Vergleich zur alleinigen Zuhilfenahme der UICC-Staging-Klassifikation bezüglich der Prognostik von Gesamtüberleben, progressionsfreiem Überleben [ 12 ] und lokoregionärer Tumorprogression [ 11 ] bei Patient*innen mit HPV-assoziierten und nicht-HPV-assoziierten Oropharynxkarzinomen.…”

Section: Ki In Der Kopf-hals-onkologieunclassified

Künstliche Intelligenz in der Hals-Nasen-Ohren-Heilkunde

Haider

Sharaf

Baumeister

et al. 2021

HNO

Self Cite

View full text Add to dashboard Cite

Zusammenfassung Hintergrund Die fortschreitende Digitalisierung ermöglicht zunehmend den Einsatz von künstlicher Intelligenz (KI). Sie wird Gesellschaft und Medizin in den nächsten Jahren maßgeblich beeinflussen. Ziel der Arbeit Darstellung des gegenwärtigen Einsatzspektrums von KI in der Hals-Nasen-Ohren-Heilkunde und Skizzierung zukünftiger Entwicklungen bei der Anwendung dieser Technologie. Material und Methoden Es erfolgte die Auswertung und Diskussion wissenschaftlicher Studien und Expertenanalysen. Ergebnisse Durch die Verwendung von KI kann der Nutzen herkömmlicher diagnostischer Werkzeuge in der Hals-Nasen-Ohren-Heilkunde gesteigert werden. Zudem kann der Einsatz dieser Technologie die chirurgische Präzision in der Kopf-Hals-Chirurgie weiter erhöhen. Schlussfolgerungen KI besitzt ein großes Potenzial zur weiteren Verbesserung diagnostischer und therapeutischer Verfahren in der Hals-Nasen-Ohren-Heilkunde. Allerdings ist die Anwendung dieser Technologie auch mit Herausforderungen verbunden, beispielsweise im Bereich des Datenschutzes.

show abstract

Prediction of post-radiotherapy locoregional progression in HPV-associated oropharyngeal squamous cell carcinoma using machine-learning analysis of baseline PET/CT radiomics

Cited by 22 publications

References 36 publications

Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features

Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features

Discrimination of Cancer Stem Cell Markers ALDH1A1, BCL11B, BMI-1, and CD44 in Different Tissues of HNSCC Patients

Künstliche Intelligenz in der Hals-Nasen-Ohren-Heilkunde

Contact Info

Product

Resources

About