Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Sep 23;11(1):4807.
doi: 10.1038/s41467-020-18497-3.

Lymph node metastasis prediction of papillary thyroid carcinoma based on transfer learning radiomics

Jinhua Yu  1   2 , Yinhui Deng  1   3 , Tongtong Liu  1 , Jin Zhou  4 , Xiaohong Jia  5 , Tianlei Xiao  1 , Shichong Zhou  4 , Jiawei Li  4 , Yi Guo  1 , Yuanyuan Wang  6   7 , Jianqiao Zhou  8 , Cai Chang  9
Affiliations

Lymph node metastasis prediction of papillary thyroid carcinoma based on transfer learning radiomics

Jinhua Yu et al. Nat Commun. .

Abstract

Non-invasive assessment of the risk of lymph node metastasis (LNM) in patients with papillary thyroid carcinoma (PTC) is of great value for the treatment option selection. The purpose of this paper is to develop a transfer learning radiomics (TLR) model for preoperative prediction of LNM in PTC patients in a multicenter, cross-machine, multi-operator scenario. Here we report the TLR model produces a stable LNM prediction. In the experiments of cross-validation and independent testing of the main cohort according to diagnostic time, machine, and operator, the TLR achieves an average area under the curve (AUC) of 0.90. In the other two independent cohorts, TLR also achieves 0.93 AUC, and this performance is statistically better than the other three methods according to Delong test. Decision curve analysis also proves that the TLR model brings more benefit to PTC patients than other methods.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Comparison of ROC curves of LNM prediction in the main cohort by four models.
a ROC curves in the cross-validation set, b ROC in testing set, and c decision curves in the testing set. (SM statistical model, RM traditional radiomics model, NTLR nontransfer learning radiomics, TLR transfer learning radiomics).
Fig. 2
Fig. 2. Comparison of ROC curves when data from different machines and different doctors are used as independent test sets.
a shows ROC curves when data from different machines, b from different doctors.
Fig. 3
Fig. 3. Comparison of ROC curves and decision curves of LNM prediction in the two independent testing cohort by four models.
a, c ROC curves in the independent testing set 1 and 2, respectively. b, d Decision curves in the independent testing set 1 and 2, respectively. (SM statistical model, RM traditional radiomics model, NTLR nontransfer learning radiomics, TLR transfer learning radiomics).
Fig. 4
Fig. 4. Visualization of network features of 20 cases with and without LNM, respectively.
The left side shows the network features of 20 cases without LNM, and right side shows 20 cases with LNM.
Fig. 5
Fig. 5. Process of the patient enrollment for main cohort and two independent testing sets.
The left side shows the patient enrollment process of the main cohort, the middle is the one of the independent testing set1, and the right side is the one of the independent testing set2.
Fig. 6
Fig. 6. Structure of transfer learning model and illustrations of middle layer output of an LNM positive case and a negative case.
The left side shows the structure of our model, and the right side shows the middel layer output of two cases.
See this image and copyright information in PMC

References

    1. Bray F, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 2018;68:394–424. doi: 10.3322/caac.21492. - DOI - PubMed
    1. Kitahara, C. M., Schneider, A. B. & Brenner, A. V. Thyroid cancer. In Cancer Epidemiology and Prevention 4th edn, 839–860 (Oxford University Press, New York, 2018).
    1. Lim H, Devesa SS, Sosa JA, Check D, Kitahara CM. Trends in thyroid cancer incidence and mortality in the United States, 1974-2013. JAMA. 2017;317:1338–1348. doi: 10.1001/jama.2017.2719. - DOI - PMC - PubMed
    1. Oler G, et al. Gene expression profiling of papillary thyroid carcinoma identifies transcripts correlated with BRAF mutational status and lymph node metastasis. Clin. Cancer Res. 2008;14:4735–4742. doi: 10.1158/1078-0432.CCR-07-4372. - DOI - PMC - PubMed
    1. Haugen BR, et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2016;26:1–133. doi: 10.1089/thy.2015.0020. - DOI - PMC - PubMed

Publication types