Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept

Morteza Salehjahromi; Tatiana V. Karpinets; Sheeba J. Sujit; Mohamed Qayati; Pingjun Chen; Muhammad Aminu; Maliazurina B. Saad; Rukhmini Bandyopadhyay; Lingzhi Hong; Ajay Sheshadri; Julie Lin; Mara B. Antonoff; Boris Sepesi; Edwin J. Ostrin; Iakovos Toumazis; Peng Huang; Chao Cheng; Tina Cascone; Natalie I. Vokes; Carmen Behrens; Jeffrey H. Siewerdsen; John D. Hazle; Joe Y. Chang; Jianhua Zhang; Yang Lu; Myrna C.B. Godoy; Caroline Chung; David Jaffray; Ignacio Wistuba; J. Jack Lee; Ara A. Vaporciyan; Don L. Gibbons; Gregory Gladish; John V. Heymach; Carol C. Wu; Jianjun Zhang; Jia Wu

doi:10.1016/j.xcrm.2024.101463

Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept

Morteza Salehjahromi, Tatiana V. Karpinets, Sheeba J. Sujit, Mohamed Qayati, Pingjun Chen, Muhammad Aminu, Maliazurina B. Saad, Rukhmini Bandyopadhyay, Lingzhi Hong, Ajay Sheshadri, Julie Lin, Mara B. Antonoff, Boris Sepesi, Edwin J. Ostrin, Iakovos Toumazis, Peng Huang, Chao Cheng, Tina Cascone, Natalie I. Vokes, Carmen BehrensJeffrey H. Siewerdsen, John D. Hazle, Joe Y. Chang, Jianhua Zhang, Yang Lu, Myrna C.B. Godoy, Caroline Chung, David Jaffray, Ignacio Wistuba, J. Jack Lee, Ara A. Vaporciyan, Don L. Gibbons, Gregory Gladish, John V. Heymach, Carol C. Wu, Jianjun Zhang, Jia Wu

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

[¹⁸F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.

Original language	English (US)
Article number	101463
Journal	Cell Reports Medicine
Volume	5
Issue number	3
DOIs	https://doi.org/10.1016/j.xcrm.2024.101463
State	Published - Mar 19 2024

Keywords

Turing test
generative AI
generative adversarial network
image synthesis
lung cancer
prognostic marker
radiogenomics
radiomics
screening
synthetic PET

ASJC Scopus subject areas

General Biochemistry, Genetics and Molecular Biology

Access to Document

10.1016/j.xcrm.2024.101463

Cite this

Salehjahromi, M., Karpinets, T. V., Sujit, S. J., Qayati, M., Chen, P., Aminu, M., Saad, M. B., Bandyopadhyay, R., Hong, L., Sheshadri, A., Lin, J., Antonoff, M. B., Sepesi, B., Ostrin, E. J., Toumazis, I., Huang, P., Cheng, C., Cascone, T., Vokes, N. I., ... Wu, J. (2024). Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept. Cell Reports Medicine, 5(3), Article 101463. https://doi.org/10.1016/j.xcrm.2024.101463

Salehjahromi, M , Karpinets, TV , Sujit, SJ, Qayati, M, Chen, P, Aminu, M, Saad, MB, Bandyopadhyay, R, Hong, L , Sheshadri, A , Lin, J , Antonoff, MB, Sepesi, B, Ostrin, EJ , Toumazis, I, Huang, P, Cheng, C, Cascone, T , Vokes, NI , Behrens, C , Siewerdsen, JH , Hazle, JD , Chang, JY, Zhang, J, Lu, Y , Godoy, MCB , Chung, C , Jaffray, D , Wistuba, I , Lee, JJ , Vaporciyan, AA , Gibbons, DL , Gladish, G , Heymach, JV , Wu, CC , Zhang, J & Wu, J 2024, 'Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept', Cell Reports Medicine, vol. 5, no. 3, 101463. https://doi.org/10.1016/j.xcrm.2024.101463

@article{32661a146e484dc59ede9eba49e22a01,

title = "Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept",

abstract = "[18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.",

keywords = "Turing test, generative AI, generative adversarial network, image synthesis, lung cancer, prognostic marker, radiogenomics, radiomics, screening, synthetic PET",

author = "Morteza Salehjahromi and Karpinets, {Tatiana V.} and Sujit, {Sheeba J.} and Mohamed Qayati and Pingjun Chen and Muhammad Aminu and Saad, {Maliazurina B.} and Rukhmini Bandyopadhyay and Lingzhi Hong and Ajay Sheshadri and Julie Lin and Antonoff, {Mara B.} and Boris Sepesi and Ostrin, {Edwin J.} and Iakovos Toumazis and Peng Huang and Chao Cheng and Tina Cascone and Vokes, {Natalie I.} and Carmen Behrens and Siewerdsen, {Jeffrey H.} and Hazle, {John D.} and Chang, {Joe Y.} and Jianhua Zhang and Yang Lu and Godoy, {Myrna C.B.} and Caroline Chung and David Jaffray and Ignacio Wistuba and Lee, {J. Jack} and Vaporciyan, {Ara A.} and Gibbons, {Don L.} and Gregory Gladish and Heymach, {John V.} and Wu, {Carol C.} and Jianjun Zhang and Jia Wu",

note = "Publisher Copyright: {\textcopyright} 2024 The Author(s)",

year = "2024",

month = mar,

day = "19",

doi = "10.1016/j.xcrm.2024.101463",

language = "English (US)",

volume = "5",

journal = "Cell Reports Medicine",

issn = "2666-3791",

publisher = "Cell Press",

number = "3",

}

TY - JOUR

T1 - Synthetic PET from CT improves diagnosis and prognosis for lung cancer

T2 - Proof of concept

AU - Salehjahromi, Morteza

AU - Karpinets, Tatiana V.

AU - Sujit, Sheeba J.

AU - Qayati, Mohamed

AU - Chen, Pingjun

AU - Aminu, Muhammad

AU - Saad, Maliazurina B.

AU - Bandyopadhyay, Rukhmini

AU - Hong, Lingzhi

AU - Sheshadri, Ajay

AU - Lin, Julie

AU - Antonoff, Mara B.

AU - Sepesi, Boris

AU - Ostrin, Edwin J.

AU - Toumazis, Iakovos

AU - Huang, Peng

AU - Cheng, Chao

AU - Cascone, Tina

AU - Vokes, Natalie I.

AU - Behrens, Carmen

AU - Siewerdsen, Jeffrey H.

AU - Hazle, John D.

AU - Chang, Joe Y.

AU - Zhang, Jianhua

AU - Lu, Yang

AU - Godoy, Myrna C.B.

AU - Chung, Caroline

AU - Jaffray, David

AU - Wistuba, Ignacio

AU - Lee, J. Jack

AU - Vaporciyan, Ara A.

AU - Gibbons, Don L.

AU - Gladish, Gregory

AU - Heymach, John V.

AU - Wu, Carol C.

AU - Zhang, Jianjun

AU - Wu, Jia

PY - 2024/3/19

Y1 - 2024/3/19

N2 - [18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.

AB - [18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.

KW - Turing test

KW - generative AI

KW - generative adversarial network

KW - image synthesis

KW - lung cancer

KW - prognostic marker

KW - radiogenomics

KW - radiomics

KW - screening

KW - synthetic PET

UR - http://www.scopus.com/inward/record.url?scp=85187639538&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85187639538&partnerID=8YFLogxK

U2 - 10.1016/j.xcrm.2024.101463

DO - 10.1016/j.xcrm.2024.101463

M3 - Article

C2 - 38471502

AN - SCOPUS:85187639538

SN - 2666-3791

VL - 5

JO - Cell Reports Medicine

JF - Cell Reports Medicine

IS - 3

M1 - 101463

ER -

Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this