Attention Guided Lymph Node Malignancy Prediction in Head and Neck Cancer

Chen, Liyuan; Dohopolski, Michael; Zhou, Zhiguo; Wang, Kai; Wang, Rongfang; Sher, David; Wang, Jing

doi:10.1016/j.ijrobp.2021.02.004

Home/Archive/DOE OSTI

DOE OSTI · 23198519

Attention Guided Lymph Node Malignancy Prediction in Head and Neck Cancer

Chen, Liyuan·Dohopolski, Michael·Zhou, Zhiguo·Wang, Kai·Wang, Rongfang·Sher, David·Wang, Jing

View original source ↗Export metadata ↓

Abstract

Accurate lymph node (LN) malignancy classification is essential for treatment target identification in head and neck cancer (HNC) radiation therapy. Given the constraints imposed by relatively small sample sizes in real-world medical applications, to classify LN malignancy status accurately, we proposed an attention-guided classification (AGC) scheme that (1) incorporates human knowledge (ie, LN contours) into model training to guide model’s “learning” direction, alleviating the critical requirement of large training samples by deep learning approaches; and (2) does not require accurate delineation of LNs in the inference stage but can highlight the discriminative region nearby the LN, which is important for malignancy determination.

Explore related subjects

62 RADIOLOGY AND NUCLEAR MEDICINE CLASSIFICATION HEAD LYMPH NODES MACHINE LEARNING NECK NEOPLASMS RADIOTHERAPY TRAINING

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Chen, Liyuan, Dohopolski, Michael, Zhou, Zhiguo, Wang, Kai, Wang, Rongfang, Sher, David, Wang, Jing. 2021-07-15. Attention Guided Lymph Node Malignancy Prediction in Head and Neck Cancer. https://doi.org/10.1016/j.ijrobp.2021.02.004

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

DOE OSTI2026

Generator, Chelator, and Kit Optimization (GeCKO): Bringing 44TI and 43/44gSc into Radiopharmacy Practice

To inform development of separation and generator technology relevant to the accelerator (and generator) production of 43Sc, 44Sc (from 44Ti), 45Ti, and 47Sc with fundamental, experimental exploration of the speciation and coordination chemistry of Ti and Sc.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2025

Cross sections of 147–149 Sm( 6 Li,x) reactions for the production of 149 Tb for targeted alpha therapy

Terbium-149g (t 1/2 = 4.12 h) is of particular interest for targeted alpha therapy cancer treatment due to its ability to decay via both alpha and positron emission, making it a potential theranostic nuclide. Due to many challenges facing its production, there are limited facilities worldwide that have demonstrated the ability to produce this nuclide in quantities sufficient for medical research. Since the Cyclotron Institute at Texas A&M University is a specialized accelerator facility capable of accelerating a wide variety of ions, we are investigating production pathway options. One of the major challenges facing its production is the known co-production of the excited isomeric state, 149m Tb (t 1/2 = 4.1 min). However, this state does not decay to the ground state of 149g Tb, negating any potential contribution to its yield. Due to its short-half life, the cross section for the population of this state has never been measured. After calculating several potential reaction yields using predictive models, the reactions of 147–149 Sm( 6 Li,xn) 149 Tb were identified as candidates. Lithium-6 beams of varied energies between 45-65 MeV were impinged on enriched 147 Sm, 148 Sm, and 149 Sm targets at the Cyclotron Institute at Texas A&M University, and the reaction products were measured immediately following irradiation using high-purity germanium detectors, enabling detection of both 149m Tb and 149g Tb. Cross sections for all nuclides produced in sufficient activity in these reactions were also measured and reported here. We conclude that the population of 149m Tb is much preferred over population of the ground state for these 6 Li-induced reactions, and it is necessary to explore other options for 149g Tb production.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2025

Source function from two-particle correlation function through entropy-regularized Richardson-Lucy deblurring

Source functions are obtained from p – p and d – α correlation functions by applying the Richardson-Lucy (RL) deblurring to the Koonin-Pratt (KP) equation. To prevent fitting of noise in the correlation function, total-variation (TV) regularization is employed that has been effective in ordinary image restoration. TV alone cannot ensure normalization of the source functions. To ensure the latter, we propose a maximum-entropy regularized RL algorithm (MEM-RL). We outline the MEM-RL formalism and optimization strategy for the KP equation, demonstrating its effectiveness on both simulated and experimental data, including the p – p and d – α correlation functions.

62 RADIOLOGY AND NUCLEAR MEDICINE↗