Prediction of Anatomical Presentation during Radiotherapy of Nasopharyngeal Carcinoma Using GAN-LSTM for Plan Adaption Decision

PURPOSE/OBJECTIVE(S)

During treatment, patients frequently undergo anatomical changes that might result in dose degradation. Adaptive radiation therapy (ART) is now available to overcome this problem. However, this method is time-consuming, and the lack of criteria to trigger replanning prevents its widespread use. To overcome these obstacles, anatomical presentation models are necessary. In this study, we developed a novel deep-learning method to make a series of predictions for the remaining treatment course.

MATERIALS/METHODS

Total 230 nasopharyngeal carcinoma patients who received radiotherapy in 33 fractions were enrolled. The data included cone-beam computed tomography (CBCT) and planning CT images. CBCT image quality was improved to CT level using an in-house software. A generative adversarial network-long short-term memory network model was proposed, with the generation abilities of the former network and forecasting abilities of the latter. To predict the anatomical presentation for 3-6 weeks, we trained four models. The planning CT and CBCT acquired earlier were used as the input. Physicians segmented the gross target volume (GTVnx) and parotid glands on prediction and real CBCT (ground truth). Contours, dosimetry parameters, and plan adaptation decision were used to evaluate the models.

RESULTS

The table shows the overall performance of the test set (18 cases). The anatomical changes were predicted over the treatment course with a dice similarity coefficient (DSC) of 0.96, 0.90 and 0.92 and mean distance to agreement (MDA, in mm) of 0.37, 0.70, and 0.60 for GTVnx, left parotid, and right parotid, respectively. Bland-Altman analysis revealed that dosimetry parameters did not show significant difference between prediction and ground truth. The prescription coverage (%) of GTVnx, V30 of the left parotid, and V30 of the right parotid had mean absolute biases of 0.09, 1.09, and 0.27, respectively. At week 6, there were two cases that required plan adaptation, and the model effectively triggered replanning one week in advance.

CONCLUSION

We developed a framework that predicts the anatomical changes occurring in future fractions. Establishing such a framework provides a proactive approach to ART and saves clinical time by anticipating and preparing for treatment strategies in advance.

A Machine Learning Method to Predict Pathological Complete Response of Esophageal Cancer after Neoadjuvant Chemoradiotherapy with Clinicohematological Markers and MR Radiomics: A Multi-Center Study

PURPOSE/OBJECTIVE(S)

Nearly 30% of patients with local advanced esophageal cancer achieved pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT), who may benefit from organ-preservation strategy under accurate prediction of pCR. We aimed to develop and validate machine learning models based on clinicohematological markers and MR radiomics to accurately predict pCR of esophageal cancer after nCRT.

MATERIALS/METHODS

In this multi-center study, eligible patients with esophageal cancer who received baseline MR scan (T2-weighted image) and nCRT plus surgery were enrolled between September 2014 and September 2022 at institution 1 (training set) and between December 2017 and August 2021 at institution 2 (testing set). Pre-nCRT and post-nCRT blood test results were collected to calculate hematological markers. Models were constructed by machine learning based on clinicohematological markers and MR radiomics to predict pCR. Area under the curve (AUC) and cut-off analysis were used to evaluate model performances.

RESULTS

Totally 154 patients (81 in the training set and 73 in the testing set) were enrolled. The combined model integrating pre-nCRT monocyte-to-lymphocyte ratio and 6 radiomics features achieved AUC of 0.800 (95% CI 0.671-0.918) in the testing set, with sensitivity of 79.2% (95% CI 62.5%-95.8%), specificity of 83.7% (95% CI 73.5%-93.9%), positive predictive value of 76.0% (95% CI 62.5%-90.0%), and negative predictive value of 89.6% (95% CI 82.0%-95.8%).

CONCLUSION

A machine learning model based on clinicohematological markers and MR radiomics to predict pCR after nCRT for patients with esophageal cancer was developed and validated, providing a novel tool for personalized treatment. It is necessary to further validate in more large datasets.

More Efficient Auto-Segmentation Framework Using Patient-Specific Information for CBCT-Guided Adaptive Radiotherapy

PURPOSE/OBJECTIVE(S)

Accurate and fast delineation of regions of interest (ROIs) is critical for online adaptive radiotherapy (ART). Due to the noise, artifacts and low contrast of soft tissue on cone-beam computed tomography (CBCT), the CBCT-guided ART (CBCTgART) needs an effective tool to assist the segmentation of ROIs. This study aims to develop an efficient auto-segmentation workflow with personalized modeling for CBCTgART based on deep learning (DL) model.

MATERIALS/METHODS

Five hundred fifty-two patients with nasopharyngeal carcinoma were enrolled in this study. At the beginning, a cycle consistent adversarial network (CycleGAN) model was trained on 172 patients' CBCT/CT paired images to improve the image quality of CBCT to CT level. A generalized segmentation model was trained using the planning CT (pCT) and contour data from 530 patients. For personalized modeling, the generalized segmentation model was fine-tuned on the specific patient's pCT and contour to get the personalized model. When CBCT images were available, the trained CycleGAN model transformed the CBCT to synthetic pCT. Then the personalized auto-segmentation model generated the contour of ROIs on the synthetic pCT. We randomly selected 22 patients for model test. The proposed method (DL-personalized) was compared with other DL methods based on the same architecture of network: regular deep learning method (DL-regular), which was trained on the CBCT and corresponding contours, and generalized model in our framework (DL-generalized). So, 22 personalized, 1 generalized and 1 regular DL models were tested. The paired t-test was performed to test the significance for mean dice similarity coefficient (DSC), mean distance to agreement (MDA), and Hausdorff distance (HD) between the alternative and proposed methods.

RESULTS

Two ROIs were included: the clinical target volume (CTV) and nasopharynx gross tumor volume (GTVnx). The proposed DL-personalized model achieved better results compared with others as shown in the table. The accuracy and robustness of our proposed framework was reliable. All of p values were under 0.01, which indicated the statistically significant difference.

CONCLUSION

The proposed framework utilizing patient-specific information can improve the segmentation accuracy of ART. This method has potential to be integrated into the ART workflow to improve efficiency.

Quality Assurance in a Phase III, Multicenter, Randomized Trial of POstmastectomy radioThErapy in Node posiTive Breast Cancer with or without Internal mAmmary nodaL Irradiation (POTENTIAL): A Planning Dummy Run

PURPOSE/OBJECTIVE(S)

To report the planning dummy run results of the POstmastectomy radioThErapy in Node posiTive breast cancer with or without Internal mAmmary nodaL irradiation (POTENTIAL) trial-a multicenter, randomized, phase 3 trial-to evaluate postmastectomy radiotherapy, with or without internal mammary nodal irradiation, for patients with high-risk breast cancer.

MATERIALS/METHODS

All participating institutions were provided the contours of the dummy run case, and they generated radiotherapy (RT) plans per protocol guidelines. The plans were reviewed and feedback were provided by the quality assurance team, after which the institutions resubmitted revised plans. The information on beams arrangement, skin flash, inhomogeneity corrections, and protocol compliance was assessed both in the primary and final submission.

RESULTS

Theplans from 26 institutions were included in the analysis. A number of major deviations were found in the primary submission, such as less strict constraint on organs at risk (OARs) V5Gy, and no application of chest wall skin flash. The protocol compliance rates of the dose coverage for the planning target volume of the chest wall (PTVcw), PTV of supra/infraclavicular fossa plus axilla levels I, II, III (PTVsc+ax), and PTV of the IMN region (PTVim) were all significantly improved in the final submission compared with those in the primary submission, which were 96.2% vs. 69.2%, 100% vs. 76.9%, and 88.4% vs. 53.8, respectively. For OARs, the protocol compliance rates of heart Dmean, left anterior descending coronary artery V40Gy, ipsilateral lung V5Gy, and stomach V5Gy were significantly improved.

CONCLUSION

All major deviations were corrected and protocol compliance was significantly improved and of high level in the final submission. Moreover, the variations were reduced. Therefore, a planning dummy run was essential to guarantee good RT plan quality and inter-institutional consistency for multicenter trials.

Sex-specific effect of perfluoroalkyl substances exposure on liver and thyroid function biomarkers: A mixture approach

Although studies have investigated the effects of perfluoroalkyl substances (PFASs) on liver and thyroid function, little is known about its combined and sex-specific effect. A total of 688 participants were interviewed and serum PFASs concentration was measured using liquid chromatography/mass spectrometry. Five biomarkers of liver and thyroid function (ALT, GGT, TSH, FT3 and FT4) were chosen as outcomes. A restriction cubic spline function was applied to capture the dose-response relationship between PFASs and liver enzymes and thyroid hormones. Multivariable regression and Bayesian kernel machine regression (BKMR) models were performed to assess the single and overall associations of PFASs with targeted biomarkers. Single-pollutant analyses indicated that increased PFASs concentrations were associated with elevated ALT and GGT levels. BKMR models suggested positive dose-response relationships between PFASs mixtures and ALT and GGT levels. Significant associations were only detected between several PFASs and thyroid hormones, and joint effect of PFASs mixtures on FT3 levels was found at higher concentrations. Meanwhile, sex differences were found in the associations of PFASs with ALT and GGT levels, with significant results only in males. Our findings provide epidemiological evidence for combined and sex-specific effects of PFASs on ALT and GGT levels.

Association between exposure to perfluoroalkyl substances and uric acid in Chinese adults

BACKGROUND

A growing body of evidence suggests the deleterious effects of perfluoroalkyl substances (PFASs) on kidney, but little is known on the association between PFASs joint exposure and uric acid.

METHODS

Serum PFASs concentrations were measured in 661 participants recruited from Tianjin, China using liquid chromatography/mass spectrometry. The associations of single PFASs exposure with uric acid levels and hyperuricemia were assessed using multivariable linear and logistic regression models, respectively. Restricted cubic spline models were established to investigate the dose-response relationships between PFASs concentrations and uric acid levels. Bayesian Kernel Machine Regression (BKMR) model with a hierarchical variable selection was performed to assess the joint effect of PFASs on uric acid.

RESULTS

Potassium perfluoro-1-octanesulfonate (PFOS) and perfluoro-n-octanoic acid (PFOA) were the dominated contributors with median concentrations of 16.80 ng/ml and 9.42 ng/ml, respectively. Increased PFOA concentration (per log₂-unit) was associated with elevated uric acid level (β = 0.088, 95% CI: 0.033-0.143) and higher risk of hyperuricemia (OR = 1.134, 95% CI: 1.006-1.289). Conversely, the estimated change of uric acid associated with log₂-unit increment in perfluoro-n-decanoic acid (PFDA) was -0.081 mg/dL (95% CI: -0.154, -0.009). A significant linear dose-response pattern was found between log₂-transformed PFOA concentration and uric acid level. BKMR analyses indicated a non-significant overall effect of PFASs mixture on uric acid.

CONCLUSIONS

Significant associations between PFOA and PFDA and uric acid, and between PFOA and hyperuricemia were found in the single-pollutant models, but the joint effect of PFASs mixture on uric acid was not observed in the BKMR model, which provided new insights in regulation policies and risk assessment of PFASs.

Short‐term outcomes in patients with lung cancer‐associated acute ischemic stroke

Background

To investigate the independent risk factors of poor short‐term outcomes in patients with lung cancer‐associated acute ischemic stroke (LCAIS) and use them to develop an index of prognosis LCAIS (pLCAIS) which could help clinicians identify patients at high risk for poor short‐term outcomes.

Methods

We retrospectively enrolled patients with lung cancer‐associated acute ischemic stroke and employed the 90D modified Rankin cale (mRS) to divide them into two groups: good outcomes (score 0–2) and poor outcomes (score 3–6). Propensity score matching (PSM) was used to remove confounding factors, and multivariable logistic regression analysis was used to analyze the independent risk factors of pLCAIS. The receiver operating characteristic (ROC) and area under the ROC curve (AUC) developed a multiple model combining the independent risk factors of pLCAIS.

Results

A total of 172 patients were included: 67 (38.9%) with good outcomes and 105 (61.1%) with poor outcomes. After using PSM, there were 33 cases in each group. The results showed that patients with poor short‐term outcomes were significantly higher in D‐dimer (OR = 1.001, 95% CI: 1.000–1.002, p = 0.048), CRP (OR = 1.078, 95% CI: 1.008–1.153, p = 0.028), and neutrophil count (OR = 14.673, 95% CI: 1.802–19.500, p = 0.012). The ROC curve, used to assess the diagnostic ability of binary classifiers, showed that the product of these three independent risk factors showed high sensitivity and specificity.

Conclusion

In this study, we have identified three independent risk factors associated with poor short‐term outcomes in pLCAIS: higher NC, CRP, and D‐dimer levels. These findings may be helpful for clinicians in identifying poor short‐term outcomes patients.

Effect of artemisinin on improving islet function in rats fed with maternal high-fat diet

BACKGROUND

Maternal high-fat diet (HFD) is a detrimental factor in developing glucose intolerance, obesity, and islet dysfunction. However, the effect of artemisinin on maternal HFD and whether it is related to the alterations of islet function is seldom studied since artemisinin treatments not only attenuate insulin resistance (IR) and restore islet ß cell function in Diabetes mellitus type 2.

METHODS

Female rats were randomly fed a HFD (45% kcal from fat), HFD + artemisinin, or a regular chow diet (RCD) before pregnancy and during gestation. Glucose metabolism and the β cell phenotypes were assessed.

RESULTS

Maternal HFD increased islet load in female rats, proliferation of pancreatic β cells, increased insulinogen, and decreased insulin secretion response to high glucose stimulation with delayed insulin release, increased fasting glucose, and glucose area under the curve compared with the general diet group. HFD inhibited expression of Foxo1 and PAX6 in female rats. Under the effect of both HFD and pregnancy, islet load was further increased, insulinogen was further increased, and fasting insulin level and fasting glucose were higher than RCD fed general-pregnancy group. ALDH1a3 transdifferentiation and PAX6, Foxo1, and PDX1 expression were increased in islets of high-fat pregnant rats. When adding artemisinin in HFD treated pregnant rats, islet function was significantly improved.

CONCLUSIONS

Intervention with artemisinin in maternal HFD resulted in reduced islet size, decreased number of β-cells and improved islet microcirculation, insulin processing shear process, decreased insulinogen/insulin ratio, and restored islet function through increased expression of PC1/3.

The evaluation of cellular immune function in elderly patients with systemic lupus erythematosus

BACKGROUND/AIMS

To evaluate cellular immune function in systemic lupus erythematosus (SLE) patients over 60 years old, the association between antinuclear antibody (ANA) titers and the ratio of CD4+ /CD8+ was analyzed in this study. The distribution of ANAs and extractable nuclear antibodies (ENAs) in a healthy elderly population was also investigated.

METHODS

Serum ANA titers were assayed by indirect immunofluorescence (IIF) and the CD4+ /CD8+ T-cell ratio was determined by flow cytometry in 76 SLE patients and 30 healthy control individuals. IIF of cytoplasm and nuclear and nucleolar staining were performed on samples taken from 286 healthy elderly individuals. ENA levels were determined using a strip enzyme immunoassay among patients that tested positive for ANAs.

RESULTS

ANA titers were negative in the 30 control individuals, but were positive in the 76 SLE patients. Based on ANA titers, the SLE patients were stratified to low (≤ 1:320), medium (1:640 to 1:1,280), and high (≥ 1:2,560) titer groups. The average CD4+ /CD8+ ratio of the SLE group was significantly lower than that of the control group. Among the 286 healthy elderly volunteers, 59 (20.63%) tested positive for ANAs. A homogeneous pattern was present in 47.46% of those 59 patients and a granule pattern in the karyoplasm was present in 33.90%. Furthermore, of the 59 patients, ENAs immunoassay was positive in 18 (30.51%); Sjogren syndrome-related antigen A (SSA)/52 kd and Sjogren syndrome antigen B (SSB)/La were the two major antibodies.

CONCLUSION

The significantly lower CD4+ /CD8+ ratio among SLE patients over 60 years old is associated with deregulated immune responses and the development of SLE. A low ANA titer (1:160) is common in healthy elders, emphasizing the importance of considering age when determining if the evaluation of ANA titers is to be included in autoimmune disease diagnosis.

Efficacy and safety of dipeptidyl peptidase-4 inhibitors in the treatment of type 2 diabetes mellitus patients with moderate to severe renal impairment: a meta-analysis

OBJECTIVE

Dipeptidyl peptidase-4 (DPP-4) inhibitors are a new class of oral antidiabetic agents for type 2 diabetes mellitus (T2DM) patients. However, the effects and safety of DPP-4 inhibitors in T2DM patients with renal impairment (RI) remain controversial. Therefore, we conducted this meta-analysis to assess the efficacy and safety of DPP-4 inhibitors in T2DM patients with moderate to severe RI.

MATERIALS AND METHODS

The PubMed, Embase, and Web of Science database were searched for published randomized controlled trials (RCTs), which compared DPP-4 inhibitors with placebo or a control regimen. A fixed-model effect or random-effect model was used to assess the effects of DPP-4 inhibitors on T2DM patients with RI. Subgroup analysis or meta-regression analysis were performed to explore the potential sources of heterogeneity among the included studies.

RESULTS

13 RCTs with a total of 2,940 patients were included in this meta-analysis. Compared with other treatments, DPP-4 inhibitors were associated with a greater change in HbA1c level (weight mean difference (WMD)=-0.50, 95%CI: -0.61, -0.39; p<0.001), and a higher response rate of patients achieving the HbA1c goal of <7% (risk ratio (RR)=1.38, 95%CI: 1.12, 1.70; p=0.002). Subgroup analysis suggested that the reduced HbA1c was observed in all types of DPP-4 inhibitors, and in patients with moderate or severe RI, but not in those with end-stage renal disease. DPP-4 inhibitors did not significantly lower the FPG level (WMD=-0.36, 95%CI: -0.92, 0.20; p=0.204), and this was seen in all types of DPP-4 inhibitors except gemigliptin, which showed a significant reduction in FPG level. The prevalence of adverse events (RR=0.98, 95%CI: 0.94, 1.02; p=0.256) in the two groups was not significantly different, and DPP-4 inhibitors did not induce a higher rate of hypoglycemia (RR=1.31, 95%CI: 0.97, 1.77; p=0.075).

CONCLUSIONS

DPP-4 inhibitors significantly lowered HbA1c levels in T2DM patients with moderate to severe RI. And the treatment of DPP-4 inhibitors did not increase the risk of hypoglycemia and adverse events. Considering the potential limitations in this meta-analysis, more large-scale, well-conducted RCTs are needed to identify our findings.

[The relationship between the level of N-terminal pro-B-type natriuretic peptide and prognosis in patients with sepsis]

OBJECTIVE

To detect the relationship between the level of N-terminal pro-B-type natriuretic peptide (NT-proBNP) and prognosis in patients with sepsis.

METHODS

A prospective controlled study was conducted. Sixty cases, including 39 sepsis and 21 severe sepsis, were enrolled from May 2010 to October 2011 in intensive care unit (ICU) of Second Hospital of Tianjin Medical University. The serum level of NT-proBNP was determined and acute physiology and chronic health evaluation II (APACHEII) score was calculated on the 1st and 3rd day. The receiver operator characteristic curve (ROC curve) was draw. According to the 28-day prognosis, all patients were divided into the survival group (n=42) or the death group (n=18). At the same time 30 healthy people were enrolled as control group.

RESULTS

The level of NT-proBNP in the sepsis patients on the 1st and 3rd day were significantly higher than those of healthy controls (65.77±34.78 ng/L, 74.23±42.12 ng/L vs. 48.36±35.53 ng/L, P<0.05 and P<0.01). The level on 1st day of the severe sepsis group was higher than sepsis group (71.69±32.86 ng/L vs. 50.11±36. 98 ng/L, P<0.05), but there was no statistically significance on the 3rd day. The level of NT-proBNP in death group was increased gradually and significantly higher than that of survival group on the 3rd days (99.20±44.34 ng/L vs. 66.79±39.28 ng/L, P<0.05), but no difference was found on the 1st day. The APACHEII score of severe sepsis group were much higher than those of sepsis group on the 1st and 3rd day (1st day:23.92±7.57 vs. 14.87±6.50, 3rd day:19.28±8.80 vs. 10.43±7.27, both P<0.01). The APACHEII score of death group were also much higher than those of survival group on the 1st and 3rd day (1st day:26.71±6.72 vs. 18.83±7.84, 3rd day:31.11±5.06 vs. 13.80±7.27, both P<0.01). The cut point for the evaluation of sepsis prognosis were NT-proBNP≥63.5 ng/L and APACHEII score≥20, which sensitivity were 65.4% and 88.5%, and specificity were 62.5% and 69.4% respectively.

CONCLUSION

Serum NT-proBNP levels elevation imply the poor prognosis in patients with sepsis.

[A follow-up study on correlated factors for intrauterine infection of hepatitis B virus]

One hundred and twenty-two pregnant women with positive serum hepatitis B surface antigen (HBsAg) and their infants were followed-up to study the risk factors related to intrauterine infection of hepatitis B virus (HBV). Infants were immunized with three doses of hepatitis B vaccine within 24 hours after birth, one month and six months of age, respectively, and hepatitis B immunoglobulin (HBIG) was injected simultaneously with the first dose. Markers of HBV infection in pregnant women and infants were detected by enzyme linked immunosorbent assay (ELISA). Results showed that 13 infants were detected positive for HBsAg in their sera, eight of them were positive at their birth and the other five converted positive during follow-up. Simple and multiple logistic regression analyses showed that positivity of hepatitis Be antigen (HBeAg) in mothers and their threatened abortion related to intrauterine infection, with relative risks of 31.27 and 10.87, respectively.

[The preliminary study on the hepatitis B virus infection rate of peripheral blood mononuclear cell in both HBsAg and HBeAg-positive pregnant women and its role in intrauterine transmission]

OBJECTIVE

To study the hepatitis B virus (HBV) infection rate of peripheral blood mononuclear cells (PBMC) and the role of infected PBMC in the intrauterine transmission.

METHODS

Polymerase chain reaction (PCR) was using for detection of HBV-DNA in PBMC of 52 both HBsAg and HBeAg positive pregnant women just before onset of labour, and HBsAg were detected in sera of their neonates by enzyme linked immunoabsorbent assay (ELISA) within 24 hours after birth.

RESULTS

HBV infection rate of PBMC in these women was 46.2% (24/52). There was positive correlation between infection rate of PBMC and HBsAg titre of pregnant women (chi 2 = 4.59, P < 0.05), but no relationship between PBMC infection rate and neonatal HBV infection.

CONCLUSION

The possibility of intrauterine transmission of HBV by infected PBMC through transplacental passage was small.