A deep learning model to predict dose–volume histograms of organs at risk in radiotherapy treatment plans

2020 ◽  
Vol 47 (11) ◽  
pp. 5467-5481
Author(s):  
Zhiqiang Liu ◽  
Xinyuan Chen ◽  
Kuo Men ◽  
Junlin Yi ◽  
Jianrong Dai
Keyword(s):  
At Risk ◽  
Deep Learning ◽  
Organs At Risk ◽  
Learning Model ◽  
Dose Volume Histograms ◽  
Dose Volume ◽  
Treatment Plans ◽  
Deep Learning Model ◽  
Radiotherapy Treatment

scholarly journals The Dosimetric Impact of Deep Learning-Based Auto-Segmentation of Organs at Risk on Nasopharyngeal and Rectal Cancer

2021 ◽  
Author(s):  
Hongbo Guo ◽  
Jiazhou Wang ◽  
Xiang Xia ◽  
Yang Zhong ◽  
Jiayuan Peng ◽  
...  
Keyword(s):  
At Risk ◽  
Rectal Cancer ◽  
Deep Learning ◽  
Organs At Risk ◽  
Dice Similarity Coefficient ◽  
Pass Rates ◽  
Temporal Lobes ◽  
Dose Volume ◽  
Treatment Plans ◽  
Dosimetric Analysis

Abstract PurposeTo investigate the dosimetric impact of deep learning-based auto-segmentation of organs at risk (OARs) on nasopharyngeal and rectal cancer.Methods and MaterialsTwenty patients, including ten nasopharyngeal carcinoma (NPC) patients and ten rectal cancer patients, who received radiotherapy in our department were enrolled in this study. Two deep learning-based auto-segmentation systems, including an in-house developed system (FD) and a commercial product (UIH), were used to generate two auto-segmented OARs sets (OAR_FD and OAR_UIH). Treatment plans based on auto-segmented OARs and following our clinical requirements were generated for each patient on each OARs set (Plan_FD and Plan_UIH). Geometric metrics (Hausdorff distance (HD), mean distance to agreement (MDA), the Dice similarity coefficient (DICE) and the Jaccard index) were calculated for geometric evaluation. The dosimetric impact was evaluated by comparing Plan_FD and Plan_UIH to original clinically approved plans (Plan_Manual) with dose-volume indices and 3D gamma pass rates. Spearman’s correlation analysis was performed to investigate the correlation between dosimetric deviation and geometric metrics.ResultsFD and UIH could provide similar geometric performance in parotids, temporal lobes, lens, and eyes (DICE, p > 0.05). OAR_FD had better geometric performance in the optic nerves, oral cavity, larynx, and femoral heads (DICE, p < 0.05). OAR_UIH had better geometric performance in the bladder (DICE, p < 0.05). In dosimetric analysis, both Plan_FD and Plan_UIH had nonsignificant dosimetric differences compared to Plan_Manual for most PTV and OARs dose-volume indices. The only significant dosimetric difference was the Dmax of the left temporal lobe for Plan_FD vs. Plan_Manual (p = 0.05). Only one significant correlation was found between the mean dose of the femoral head and its HD index (R = 0.4, p = 0.01).ConclusionsDeep learning-based OARs auto-segmentation for NPC and rectal cancer has a nonsignificant impact on most PTV and OARs dose-volume indices. Correlations between the auto-segmentation geometric metric and dosimetric difference were not observed for most OARs.

scholarly journals The dosimetric impact of deep learning-based auto-segmentation of organs at risk on nasopharyngeal and rectal cancer

2021 ◽  
Vol 16 (1) ◽  
Author(s):  
Hongbo Guo ◽  
Jiazhou Wang ◽  
Xiang Xia ◽  
Yang Zhong ◽  
Jiayuan Peng ◽  
...  
Keyword(s):  
At Risk ◽  
Rectal Cancer ◽  
Deep Learning ◽  
Organs At Risk ◽  
Dice Similarity Coefficient ◽  
Temporal Lobes ◽  
Gamma Analysis ◽  
Dose Volume ◽  
Treatment Plans ◽  
Dosimetric Analysis

Abstract Purpose To investigate the dosimetric impact of deep learning-based auto-segmentation of organs at risk (OARs) on nasopharyngeal and rectal cancer. Methods and materials Twenty patients, including ten nasopharyngeal carcinoma (NPC) patients and ten rectal cancer patients, who received radiotherapy in our department were enrolled in this study. Two deep learning-based auto-segmentation systems, including an in-house developed system (FD) and a commercial product (UIH), were used to generate two auto-segmented OARs sets (OAR_FD and OAR_UIH). Treatment plans based on auto-segmented OARs and following our clinical requirements were generated for each patient on each OARs sets (Plan_FD and Plan_UIH). Geometric metrics (Hausdorff distance (HD), mean distance to agreement (MDA), the Dice similarity coefficient (DICE) and the Jaccard index) were calculated for geometric evaluation. The dosimetric impact was evaluated by comparing Plan_FD and Plan_UIH to original clinically approved plans (Plan_Manual) with dose-volume metrics and 3D gamma analysis. Spearman’s correlation analysis was performed to investigate the correlation between dosimetric difference and geometric metrics. Results FD and UIH could provide similar geometric performance in parotids, temporal lobes, lens, and eyes (DICE, p > 0.05). OAR_FD had better geometric performance in the optic nerves, oral cavity, larynx, and femoral heads (DICE, p < 0.05). OAR_UIH had better geometric performance in the bladder (DICE, p < 0.05). In dosimetric analysis, both Plan_FD and Plan_UIH had nonsignificant dosimetric differences compared to Plan_Manual for most PTV and OARs dose-volume metrics. The only significant dosimetric difference was the max dose of the left temporal lobe for Plan_FD vs. Plan_Manual (p = 0.05). Only one significant correlation was found between the mean dose of the femoral head and its HD index (R = 0.4, p = 0.01), there is no OARs showed strong correlation between its dosimetric difference and all of four geometric metrics. Conclusions Deep learning-based OARs auto-segmentation for NPC and rectal cancer has a nonsignificant impact on most PTV and OARs dose-volume metrics. Correlations between the auto-segmentation geometric metric and dosimetric difference were not observed for most OARs.

Dose volume histogram analysis for organs at risk when using 6 external beam techniques for radical prostatic irradiation

2000 ◽  
Vol 2 (1) ◽  
pp. 17-25 ◽  
Author(s):  
T. Haycocks ◽  
J. Mui ◽  
H. Alasti ◽  
C. Catton
Keyword(s):  
At Risk ◽  
Organs At Risk ◽  
Area Under The Curve ◽  
Target Volume ◽  
Dose Volume Histograms ◽  
Treatment Techniques ◽  
Dose Volume ◽  
Femoral Heads ◽  
Curve Method ◽  
Set Up

Ten patients with prostate cancer were each planned with 3 conventional and 3 conformal isocentric treatment techniques to compare the relative radiation doses to the bladder and rectal walls, and femoral head using dose volume histograms (DVH). The DVH were calculated for each organ and each technique, and the plans were ranked using the area under the curve method and also by the relative radiation dose given to specific normal tissue volumes.The results show that for the planning target volume chosen, the 4 field non-coplanar technique delivers the least dose to the bladder, the 6 field coplanar technique delivers the least dose to the rectum and the 3 field oblique technique delivers the least dose to the femoral heads. The 4-field technique with no shielding contributes the most dose to the bladder and rectum and the 6 field coplanar technique contributes the most dose to the femoral heads.No technique was shown to be optimal for all the organs at risk, but both the 6 field and 4 field non-coplanar field arrangements were shown to be superior techniques for minimising both the bladder and rectal dosage. The choice of technique will therefore depend on other factors such as the total prescribed dose, the ease of set-up and the ease of verification of isocentre reproducibility.

Evaluation of plan optimisers in prostate VMAT using the dose distribution index

2019 ◽  
Vol 18 (4) ◽  
pp. 323-328 ◽  
Author(s):  
James C. L. Chow ◽  
Runqing Jiang ◽  
Lu Xu
Keyword(s):  
At Risk ◽  
Dose Distribution ◽  
Organs At Risk ◽  
Volumetric Modulated Arc Therapy ◽  
Target Volume ◽  
Weighting Factor ◽  
Target Coverage ◽  
Dose Volume Histograms ◽  
Dose Volume ◽  
Distribution Index

AbstractPurpose:Dose distribution index (DDI) is a treatment planning evaluation parameter, reflecting dosimetric information of target coverage that can help to spare organs at risk (OARs) and remaining volume at risk (RVR). The index has been used to evaluate and compare prostate volumetric modulated arc therapy (VMAT) plans using two different plan optimisers, namely photon optimisation (PO) and its predecessor, progressive resolution optimisation (PRO).Materials and methods:Twenty prostate VMAT treatment plans were created using the PO and PRO in this retrospective study. The 6 MV photon beams and a dose prescription of 78 Gy/39 fractions were used in plans with the same dose–volume criteria for plan optimisation. Dose–volume histograms (DVHs) of the planning target volume (PTV), as well as of OARs such as the rectum, bladder, left and right femur were determined in each plan. DDIs were calculated and compared for plans created by the PO and PRO based on DVHs of the PTV and all OARs.Results:The mean DDI values were 0·784 and 0·810 for prostate VMAT plans created by the PO and PRO, respectively. It was found that the DDI of the PRO plan was about 3·3% larger than the PO plan, which means that the dose distribution of the target coverage and sparing of OARs in the PRO plan was slightly better. Changing the weighting factors in different OARs would vary the DDI value by ∼7%. However, for plan comparison based on the same set of dose–volume criteria, the effect of weighting factor can be neglected because they were the same in the PO and PRO.Conclusions:Based on the very similar DDI values calculated from the PO and PRO plans, with the DDI value in the PRO plan slightly larger than that of the PO, it may be concluded that the PRO can create a prostate VMAT plan with slightly better dose distribution regarding the target coverage and sparing of OARs. Moreover, we found that the DDI is a simple and comprehensive dose–volume parameter for plan evaluation considering the target, OARs and RVR.

Predicting dose-volume histograms for organs-at-risk in IMRT planning

2012 ◽  
Vol 39 (12) ◽  
pp. 7446-7461 ◽  
Author(s):  
Lindsey M. Appenzoller ◽  
Jeff M. Michalski ◽  
Wade L. Thorstad ◽  
Sasa Mutic ◽  
Kevin L. Moore
Keyword(s):  
At Risk ◽  
Organs At Risk ◽  
Dose Volume Histograms ◽  
Dose Volume ◽  
Imrt Planning

Volumetric modulated arc therapy versus 3D conformal radiotherapy for selected childhood neoplasms

2014 ◽  
Vol 14 (1) ◽  
pp. 35-42 ◽  
Author(s):  
Marzanna Chojnacka ◽  
Anna Zygmuntowicz-Piętka ◽  
Anna Semaniak ◽  
Katarzyna Pędziwiatr ◽  
Ryszard Dąbrowski ◽  
...  
Keyword(s):  
At Risk ◽  
Organs At Risk ◽  
Volumetric Modulated Arc Therapy ◽  
Conformal Radiotherapy ◽  
Target Volume ◽  
High Dose ◽  
Dose Volume Histograms ◽  
Dose Volume ◽  
Arc Therapy ◽  
Childhood Neoplasms

AbstractAimThe comparative study of the plan quality between volumetric modulated arc therapy (VMAT) and 3D conformal therapy (3DCRT) for the treatment of selected representative childhood neoplasms was performed.Materials and methodsDuring the year 2013, 44 children with neoplasms were irradiated using VMAT. The 3DCRT plans were created retrospectively and compared with the VMAT plans for four tumour locations. The conformity parameters, dose volume histograms for target volume and organs at risk, number of monitor units and time used to deliver the single fraction were evaluated and compared for each plan. Additionally, for patients with brain tumour the comparison of different arcs configuration was made.ResultsVMAT modality presented the superiority over older conformal methods with regard to the improvement in the dose conformity and normal tissue sparing. The noncoplanar arcs arrangement was beneficial in the decrease of high-dose volume and the protection of the organs at risk located oppositely to the target volume.FindingsVMAT could be preferred technique for treating childhood neoplasms, especially when the complex-shaped target volume is localised close to the critical structures. The noncoplanar arcs arrangement could be the method of choice in the reirradiated patients and in these with laterally located brain tumours.

Inter-operator variability in organs at risk delineation: Their effects on dose-volume histograms

2013 ◽  
Vol 29 ◽  
pp. e39
Author(s):  
D. Llanas ◽  
J. Bézin ◽  
A. Ben Abdennebi ◽  
C. Veres ◽  
D. Lefkopoulos ◽  
...  
Keyword(s):  
At Risk ◽  
Organs At Risk ◽  
Dose Volume Histograms ◽  
Dose Volume

scholarly journals A deep-learning model to continuously predict severe acute kidney injury based on urine output changes in critically ill patients

2021 ◽  
Author(s):  
Francesca Alfieri ◽  
Andrea Ancona ◽  
Giovanni Tripepi ◽  
Dario Crosetto ◽  
Vincenzo Randazzo ◽  
...  
Keyword(s):  
At Risk ◽  
Acute Kidney Injury ◽  
Logistic Regression ◽  
Regression Analysis ◽  
Deep Learning ◽  
Logistic Regression Analysis ◽  
Urine Output ◽  
Kidney Injury ◽  
Learning Model ◽  
Deep Learning Model

Abstract Background Acute Kidney Injury (AKI), a frequent complication of pateints in the Intensive Care Unit (ICU), is associated with a high mortality rate. Early prediction of AKI is essential in order to trigger the use of preventive care actions. Methods The aim of this study was to ascertain the accuracy of two mathematical analysis models in obtaining a predictive score for AKI development. A deep learning model based on a urine output trends was compared with a logistic regression analysis for AKI prediction in stages 2 and 3 (defined as the simultaneous increase of serum creatinine and decrease of urine output, according to  the Acute Kidney Injury Network (AKIN) guidelines). Two retrospective datasets including 35,573 ICU patients were analyzed. Urine output data were used to train and test the logistic regression and the deep learning model. Results The deep learning model defined an area under the curve (AUC) of 0.89 (± 0.01), sensitivity = 0.8 and specificity = 0.84, which was higher than the logistic regression analysis. The deep learning model was able to predict 88% of AKI cases more than 12 h before their onset: for every 6 patients identified as being at risk of AKI by the deep learning model, 5 experienced the event. On the contrary, for every 12 patients not considered to be at risk by the model, 2 developed AKI. Conclusion In conclusion, by using urine output trends, deep learning analysis was able to predict AKI episodes more than 12 h in advance, and with a higher accuracy than the classical urine output thresholds. We suggest that this algorithm could be integrated in the ICU setting to better manage, and potentially prevent, AKI episodes. Graphic abstract

Sign in / Sign up

Export Citation Format

Share Document