Zobraziť v EDS

Deep-learning-based Attenuation Correction for 68Ga-DOTATATE Whole-body PET Imaging: A Dual-center Clinical Study.

Uložené v:
Podrobná bibliografia
Názov: Deep-learning-based Attenuation Correction for 68Ga-DOTATATE Whole-body PET Imaging: A Dual-center Clinical Study.
Alternate Title: 68Ga-DOTATATE Tüm Vücut PET Görüntüleme için Derin Öğrenme Tabanlı Atenüasyon Düzeltmesi: Çift Merkezli Bir Klinik Çalışma. (Turkish)
Autori: Lord, Mahsa Sobhi, Islamian, Jalil Pirayesh, Seyyedi, Negisa, Samimi, Rezvan, Farzanehfar, Saeed, Shahrbabk, Mahsa, Sheikhzadeh, Peyman
Zdroj: Molecular Imaging & Radionuclide Therapy; 2024, Vol. 33 Issue 3, p138-146, 9p
Predmety: X-ray imaging, STANDARD deviations, DEEP learning, DIAGNOSTIC imaging, SIGNAL-to-noise ratio
Abstract (English): Objectives: Attenuation correction is a critical phenomenon in quantitative positron emission tomography (PET) imaging with its own special challenges. However, computed tomography (CT) modality which is used for attenuation correction and anatomical localization increases patient radiation dose. This study was aimed to develop a deep learning model for attenuation correction of whole-body 68Ga-DOTATATE PET images. Methods: Non-attenuation-corrected and computed tomography-based attenuation-corrected (CTAC) whole-body 68Ga-DOTATATE PET images of 118 patients from two different imaging centers were used. We implemented a residual deep learning model using the NiftyNet framework. The model was trained four times and evaluated six times using the test data from the centers. The quality of the synthesized PET images was compared with the PET-CTAC images using different evaluation metrics, including the peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), mean square error (MSE), and root mean square error (RMSE). Results: Quantitative analysis of four network training sessions and six evaluations revealed the highest and lowest PSNR values as (52.86±6.6) and (47.96±5.09), respectively. Similarly, the highest and lowest SSIM values were obtained (0.99±0.003) and (0.97±0.01), respectively. Additionally, the highest and lowest RMSE and MSE values fell within the ranges of (0.0117±0.003), (0.0015±0.000103), and (0.01072±0.002), (0.000121±5.07xe-5), respectively. The study found that using datasets from the same center resulted in the highest PSNR, while using datasets from different centers led to lower PSNR and SSIM values. In addition, scenarios involving datasets from both centers achieved the best SSIM and the lowest MSE and RMSE. Conclusion: Acceptable accuracy of attenuation correction on 68Ga-DOTATATE PET images using a deep learning model could potentially eliminate the need for additional X-ray imaging modalities, thereby imposing a high radiation dose on the patient. [ABSTRACT FROM AUTHOR]
Abstract (Turkish): Amaç: Atenüasyon düzeltmesi, kantitatif pozitron emisyon tomografisi (PET) görüntülemede kendine özgü zorlukları olan kritik bir olgudur. Ancak atenüasyon düzeltmesi ve anatomik lokalizasyonun sağlanması amacıyla kullanılan bilgisayarlı tomografi (BT) yöntemi hastanın aldığı radyasyon dozunu artırmaktadır. Bu çalışmada, tüm vücut 68Ga-DOTATATE PET görüntülerinin atenüasyon düzeltmesi için bir derin öğrenme modelinin geliştirilmesi amaçlandı. Yöntem: İki farklı görüntüleme merkezinden 118 hastanın atenüasyon düzeltmesiz ve bilgisayarlı tomografi tabanlı atenüasyon düzeltmeli (CTAC) tüm vücut 68Ga-DOTATATE pozitron emisyon tomografi (PET) görüntüleri kullanıldı. NiftyNet çerçevesini kullanarak bir artık derin öğrenme modeli uyguladık. Model, merkezlerden alınan test verileri kullanılarak dört kez eğitildi ve altı kez değerlendirildi. Sentezlenen PET görüntülerinin kalitesi, tepe sinyal-gürültü oranı (PSNR), yapısal benzerlik indeksi (SSIM), ortalama kare hatası (MSE) ve kök ortalama kare hatası (RMSE) dahil olmak üzere farklı değerlendirme ölçümleri kullanılarak PET-CTAC görüntüleriyle karşılaştırıldı. Bulgular: Dört ağ eğitimi oturumunun ve altı değerlendirmenin kantitatif analizi ile en yüksek ve en düşük PSNR değerlerini sırasıyla (52,86±6,6) ve (47,96±5,09) olarak elde edildi. Benzer şekilde en yüksek ve en düşük SSIM değerleri sırasıyla (0,99±0,003) ve (0,97±0,01) olarak elde edildi. Ek olarak, en yüksek ve en düşük RMSE ve MSE değerleri sırasıyla (0,0117±0,003), (0,0015±0,000103) ve (0,01072±0,002), (0,000121±5,07xe-5) aralığında kaldı. Çalışma, aynı merkezden gelen veri kümelerinin kullanılmasının en yüksek PSNR değeri ile sonuçlandığını, farklı merkezlerden gelen veri kümelerinin kullanılmasının ise daha düşük PSNR ve SSIM değerleri ile sonuçlandığını buldu. Ayrıca her iki merkezden veri kümelerini içeren senaryolar ile en iyi SSIM ve en düşük MSE ve RMSE elde edildi. Sonuç: Derin öğrenme modeli kullanılarak 68Ga-DOTATATE PET görüntülerinde atenüasyon düzeltmesinin kabul edilebilir doğruluğu, potansiyel olarak, hastaya yüksek radyasyon dozu uygulanmasına neden olan ek X-ışını görüntüleme yöntemlerine olan ihtiyacı ortadan kaldırabilir. [ABSTRACT FROM AUTHOR]
Copyright of Molecular Imaging & Radionuclide Therapy is the property of Galenos Yayinevi Tic. LTD. STI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Databáza: Complementary Index
Popis
Abstrakt:Objectives: Attenuation correction is a critical phenomenon in quantitative positron emission tomography (PET) imaging with its own special challenges. However, computed tomography (CT) modality which is used for attenuation correction and anatomical localization increases patient radiation dose. This study was aimed to develop a deep learning model for attenuation correction of whole-body <sup>68</sup>Ga-DOTATATE PET images. Methods: Non-attenuation-corrected and computed tomography-based attenuation-corrected (CTAC) whole-body <sup>68</sup>Ga-DOTATATE PET images of 118 patients from two different imaging centers were used. We implemented a residual deep learning model using the NiftyNet framework. The model was trained four times and evaluated six times using the test data from the centers. The quality of the synthesized PET images was compared with the PET-CTAC images using different evaluation metrics, including the peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), mean square error (MSE), and root mean square error (RMSE). Results: Quantitative analysis of four network training sessions and six evaluations revealed the highest and lowest PSNR values as (52.86±6.6) and (47.96±5.09), respectively. Similarly, the highest and lowest SSIM values were obtained (0.99±0.003) and (0.97±0.01), respectively. Additionally, the highest and lowest RMSE and MSE values fell within the ranges of (0.0117±0.003), (0.0015±0.000103), and (0.01072±0.002), (0.000121±5.07xe<sup>-5</sup>), respectively. The study found that using datasets from the same center resulted in the highest PSNR, while using datasets from different centers led to lower PSNR and SSIM values. In addition, scenarios involving datasets from both centers achieved the best SSIM and the lowest MSE and RMSE. Conclusion: Acceptable accuracy of attenuation correction on <sup>68</sup>Ga-DOTATATE PET images using a deep learning model could potentially eliminate the need for additional X-ray imaging modalities, thereby imposing a high radiation dose on the patient. [ABSTRACT FROM AUTHOR]
ISSN:21461414
DOI:10.4274/mirt.galenos.2024.86422