Anti-HER2 therapy response assessment for guiding treatment (de-)escalation in early HER2-positive breast cancer using a novel deep learning radiomics model.

OBJECTIVES: Anti-HER2 targeted therapy significantly reduces risk of relapse in HER2 + breast cancer. New measures are needed for a precise risk stratification to guide (de-)escalation of anti-HER2 strategy. METHODS: A total of 726 HER2 + cases who received no/single/dual anti-HER2 targeted therapies were split into three respective cohorts. A deep learning model (DeepTEPP) based on preoperative breast magnetic resonance (MR) was developed. Patients were scored and categorized into low-, moderate-, and high-risk groups. Recurrence-free survival (RFS) was compared in patients with different risk groups according to the anti-HER2 treatment they received, to validate the value of DeepTEPP in predicting treatment efficacy and guiding anti-HER2 strategy. RESULTS: DeepTEPP was capable of risk stratification and guiding anti-HER2 treatment strategy: DeepTEPP-Low patients (60.5%) did not derive significant RFS benefit from trastuzumab (p = 0.144), proposing an anti-HER2 de-escalation. DeepTEPP-Moderate patients (19.8%) significantly benefited from trastuzumab (p = 0.048), but did not obtain additional improvements from pertuzumab (p = 0.125). DeepTEPP-High patients (19.7%) significantly benefited from dual HER2 blockade (p = 0.045), suggesting an anti-HER2 escalation. CONCLUSIONS: DeepTEPP represents a pioneering MR-based deep learning model that enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thereby providing valuable guidance for anti-HER2 (de-)escalation strategies. DeepTEPP provides an important reference for choosing the appropriate individualized treatment in HER2 + breast cancer patients, warranting prospective validation. CLINICAL RELEVANCE STATEMENT: We built an MR-based deep learning model DeepTEPP, which enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thus guiding anti-HER2 (de-)escalation strategies in early HER2-positive breast cancer patients. KEY POINTS: • DeepTEPP is able to predict anti-HER2 effectiveness and to guide treatment (de-)escalation. • DeepTEPP demonstrated an impressive prognostic efficacy for recurrence-free survival and overall survival. • To our knowledge, this is one of the very few, also the largest study to test the efficacy of a deep learning model extracted from breast MR images on HER2-positive breast cancer survival and anti-HER2 therapy effectiveness prediction.

Automatically optimized radiomics modeling system for small gastric submucosal tumor (<2 cm) discrimination based on EUS images.

BACKGROUND AND AIMS: The clinical management of small gastric submucosal tumors (SMTs) (<2 cm) faces a non-negligible challenge because of the lack of guideline consensus and effective diagnostic tools. This article develops an automatically optimized radiomics modeling system (AORMS) based on EUS images to diagnose and evaluate SMTs. METHODS: A total of 205 patients with EUS images of small gastric SMTs (<2 cm) were retrospectively enrolled in the development phase of AORMS for the diagnosis and the risk stratification of GI stromal tumor (GIST). A total of 178 patients with images from different centers were prospectively enrolled in the independent testing phase. The performance of AORMS was compared to that of endoscopists in the development set and evaluated in the independent testing set. RESULTS: AORMS demonstrated an area under the curve (AUC) of 0.762 for the diagnosis of GIST and 0.734 for the risk stratification of GIST, respectively. In the independent testing set, AORMS achieved an AUC of 0.770 and 0.750 for the diagnosis and risk stratification of small GISTs, respectively. In comparison, the AUCs of 5 experienced endoscopists ranged from 0.501 to 0.608 for diagnosing GIST and from 0.562 to 0.748 for risk stratification. AORMS outperformed experienced endoscopists by more than 20% in diagnosing GIST. CONCLUSIONS: AORMS implements automatic parameter selection, which enhances its robustness and clinical applicability. It has demonstrated good performance in the diagnosis and risk stratification of GISTs, which could aid endoscopists in the diagnosis of small gastric SMTs (<2 cm).

Risk Factors and Prediction Nomogram of Cognitive Frailty with Diabetes in the Elderly.

Purpose: Diabetes is a well-recognized risk factor for cognitive frailty. This study aimed to investigate the influencing factors of cognitive frailty in elderly patients with diabetes and develop a nomogram for its assessment. Methods: We collected the clinical data of diabetic patients aged 60 years or older and the patients were divided into training and validation cohorts at a ratio of 7:3. In the training cohort, logistic regression was used to screen out the influencing factors of cognitive frailty in elderly diabetic patients, and a risk prediction model and nomogram were constructed and verified in the validation cohort. The performance of the model was evaluated using various measures, including the area under the receiver operating characteristic curve, calibration curve, Hosmer-Lemeshow test and decision curve analysis. Results: A total of 315 elderly diabetic patients were included, of which 87 (27.6%) patients had cognitive frailty. Age, albumin levels, calf circumference, duration of diabetes, intellectual activity, and depressive state were identified as independent risk factors for cognitive frailty in older patients with diabetes (P < 0.05). The training cohort and validation cohort demonstrated area under curve (AUC) values of 0.866 and 0.821, respectively. Conclusion: Older patients with diabetes have a higher prevalence of cognitive frailty. The nomogram model exhibited satisfactory calibration and identification, providing a reliable tool for assessing the risk of cognitive frailty in individuals with diabetes.

Post-operative gamma knife radiosurgery for WHO grade I intracranial meningiomas: A single-center, retrospective study.

Objective: To explore the results of the Gamma Knife radiosurgery (GKRS) for World Health Organization (WHO) grade I intracranial meningiomas after surgical resection. Methods: A total of 130 patients who were pathologically diagnosed as having WHO grade I meningiomas and who underwent post-operative GKRS were retrospectively reviewed in a single center. Results: Of the 130 patients, 51 patients (39.2%) presented with radiological tumor progression with a median follow-up time of 79.7 months (ranging from 24.0 to 291.3 months). The median time to radiological tumor progression was 73.4 months (ranging from 21.4 to 285.3 months), whereas 1-, 3-, 5-, and 10-year radiological progression-free survival (PFS) was 100, 90, 78, and 47%, respectively. Moreover, 36 patients (27.7%) presented with clinical tumor progression. Clinical PFS at 1, 3, 5, and 10 years was 96, 91, 84, and 67%, respectively. After GKRS, 25 patients (19.2%) developed adverse effects, including radiation-induced edema (n = 22). In a multivariate analysis, a tumor volume of ≥10 ml and falx/parasagittal/convexity/intraventricular location were significantly associated with radiological PFS [hazard ratio (HR) = 1.841, 95% confidence interval (CI) = 1.018-3.331, p = 0.044; HR = 1.761, 95% CI = 1.008-3.077, p = 0.047]. In a multivariate analysis, a tumor volume of ≥10 ml was associated with radiation-induced edema (HR = 2.418, 95% CI = 1.014-5.771, p = 0.047). Of patients who presented with radiological tumor progression, nine were diagnosed with malignant transformation. The median time to malignant transformation was 111.7 months (ranging from 35.0 to 177.2 months). Clinical PFS after repeat GKRS was 49 and 20% at 3 and 5 years, respectively. Secondary WHO grade II meningiomas were significantly associated with a shorter PFS (p = 0.026). Conclusions: Post-operative GKRS is a safe and effective treatment for WHO grade I intracranial meningiomas. Large tumor volume and falx/parasagittal/convexity/intraventricular location were associated with radiological tumor progression. Malignant transformation was one of the main cause of tumor progression in WHO grade I meningiomas after GKRS.

Virtual elastography ultrasound via generative adversarial network for breast cancer diagnosis.

Elastography ultrasound (EUS) imaging is a vital ultrasound imaging modality. The current use of EUS faces many challenges, such as vulnerability to subjective manipulation, echo signal attenuation, and unknown risks of elastic pressure in certain delicate tissues. The hardware requirement of EUS also hinders the trend of miniaturization of ultrasound equipment. Here we show a cost-efficient solution by designing a deep neural network to synthesize virtual EUS (V-EUS) from conventional B-mode images. A total of 4580 breast tumor cases were collected from 15 medical centers, including a main cohort with 2501 cases for model establishment, an external dataset with 1730 cases and a portable dataset with 349 cases for testing. In the task of differentiating benign and malignant breast tumors, there is no significant difference between V-EUS and real EUS on high-end ultrasound, while the diagnostic performance of pocket-sized ultrasound can be improved by about 5% after V-EUS is equipped.

A deep learning model designed for Raman spectroscopy with a novel hyperparameter optimization method.

Raman spectroscopy is a spectroscopic technique typically used to determine vibrational modes of molecules and is commonly used in chemistry to provide a structural fingerprint by which molecules can be identified. With the help of deep learning, Raman spectroscopy can be analyzed more efficiently and thus provide more accurate molecular information. However, no general neural network is designed for one-dimensional Raman spectral data so far. Furthermore, different combinations of hyperparameters of neural networks lead to results with significant differences, so the optimization of hyperparameters is a crucial issue in deep learning modeling. In this work, we propose a deep learning model designed for Raman spectral data and a hyperparameter optimization method to achieve its best performance, i.e., a method based on the simulated annealing algorithm to optimize the hyperparameters of the model. The proposed model and optimization method have been fully validated in a glioma Raman spectroscopy dataset. Compared with other published methods including linear regression, support vector regression, long short-term memory, VGG and ResNet, the mean squared error is reduced by 0.1557 while the coefficient determination is increased by 0.1195 on average.

Primary versus postoperative gamma knife radiosurgery for intracranial benign meningiomas: a matched cohort retrospective study.

OBJECTIVE: The aims of this study were to investigate the long-term outcomes of primary versus postoperative Gamma Knife radiosurgery (GKRS) for benign meningiomas. METHODS: Three hundred and forty meningioma patients underwent GKRS were retrospectively reviewed. Patients in the postoperative GKRS group were matched to those in the primary GKRS group, in a 1:1 ratio. RESULTS: The study consisted of 122 patients, including primary (n = 61) and postoperative (n = 61) GKRS group. Thirty-four patients (27.9%) occurred radiological progression after a median follow-up of 72.5 (range, 24.2-254.5) months. The median time to radiological progression was 85.1 (range, 20.7-205.1) months. The radiological progression-free survival (PFS) was 100%, 93%, 87%, and 49%, at 1, 3, 5, and 10 years respectively. Thirty-one patients (25.4%) occurred clinical progression. The clinical PFS was 92%, 89%, 84%, and 60%, at 1, 3, 5, and 10 years. In combined group, only max diameter ≥ 50 mm was associated with radiological (p = 0.020) and clinical PFS (hazard ratio [HR] = 2.896, 95% confidence interval [CI] = 1.280-6.553, p = 0.011). Twenty-five patients (20.5%) developed GKRS related adverse effects, including radiation-induced edema (n = 21). Non-skull base tumors (HR = 3.611, 95% CI = 1.489-8.760, p = 0.005) and preexisting peritumoral edema (HR = 3.571, 95% CI = 1.167-10.929, p = 0.026) were significantly related to radiation-induced edema in combined group. There was no significant difference in radiological PFS (p = 0.403), clinical PFS (p = 0.336), and GKRS related adverse effects (p = 0.138) between primary and postoperative GKRS groups. CONCLUSIONS: Primary GKRS could provide similar radiological and clinical outcomes, as well as similar complication rate compared with postoperative GKRS. For selective benign meningioma patients (asymptomatic or mildly symptomatic tumors; unfavorable locations for surgical resection; comorbidities or an advanced age), GKRS could be an alternative primary treatment.

Intelligent SERS Navigation System Guiding Brain Tumor Surgery by Intraoperatively Delineating the Metabolic Acidosis.

Surgeons face challenges in intraoperatively defining margin of brain tumors due to its infiltrative nature. Extracellular acidosis caused by metabolic reprogramming of cancer cells is a reliable marker for tumor infiltrative regions. Although the acidic margin-guided surgery shows promise in improving surgical prognosis, its clinical transition is delayed by having the exogenous probes approved by the drug supervision authority. Here, an intelligent surface-enhanced Raman scattering (SERS) navigation system delineating glioma acidic margins without administration of exogenous probes is reported. With assistance of this system, the metabolites at the tumor cutting edges can be nondestructively transferred within a water droplet to a SERS chip with pH sensitivity. Homemade deep learning model automatically processes the Raman spectra collected from the SERS chip and delineates the pH map of tumor resection bed with increased speed. Acidity correlated cancer cell density and proliferation level are demonstrated in tumor cutting edges of animal models and excised tissues from glioma patients. The overall survival of animal models post the SERS system guided surgery is significantly increased in comparison to the conventional strategy used in clinical practice. This SERS system holds the promise in accelerating clinical transition of acidic margin-guided surgery for solid tumors with infiltrative nature.

Automatic Detection of Gastric Wall Structure Based on Oral Contrast-Enhanced Ultrasound and Its Application on Tumor Screening.

Gastric cancer is the second most lethal type of malignant tumor in the world. Early diagnosis of gastric cancer can reduce the transformation to advanced cancer and improve the early treatment rate. As a cheap, real-time, non-invasive examination method, oral contrast-enhanced ultrasonography (OCUS) is a more acceptable way to diagnose gastric cancer than interventional diagnostic methods such as gastroscopy. In this paper, we proposed a new method for the diagnosis of gastric diseases by automatically analyzing the hierarchical structure of gastric wall in gastric ultrasound images, which is helpful to quantify the diagnosis information of gastric diseases and is a useful attempt for early screening of gastric cancer. We designed a gastric wall detection network based on U-net. On this basis, anisotropic diffusion technology was used to extract the layered structure of the gastric wall. A simple and useful gastric cancer screening model was obtained by calculating and counting the thickness of the five-layer structure of the gastric wall. The experimental results showed that our model can accurately identify the gastric wall, and it was found that the layered parameters of abnormal gastric wall is significantly different from that of normal gastric wall. For the screening of gastric disease, a statistical model based on gastric wall stratification can give a screening accuracy of 95% with AUC of 0.92.

Characteristic of Tumor Regrowth After Gamma Knife Radiosurgery and Outcomes of Repeat Gamma Knife Radiosurgery in Nonfunctioning Pituitary Adenomas.

OBJECTIVE: This study aimed to report the characteristic of tumor regrowth after gamma knife radiosurgery (GKRS) and outcomes of repeat GKRS in nonfunctioning pituitary adenomas (NFPAs). DESIGN AND METHODS: This retrospective study consisted of 369 NFPA patients treated with GKRS. The median age was 45.2 (range, 7.2-84.0) years. The median tumor volume was 3.5 (range, 0.1-44.3) cm3. RESULTS: Twenty-four patients (6.5%) were confirmed as regrowth after GKRS. The regrowth-free survivals were 100%, 98%, 97%, 86% and 77% at 1, 3, 5, 10 and 15 year, respectively. In multivariate analysis, parasellar invasion and margin dose (<12 Gy) were associated with tumor regrowth (hazard ratio [HR] = 3.125, 95% confidence interval [CI] = 1.318-7.410, p = 0.010 and HR = 3.359, 95% CI = 1.347-8.379, p = 0.009, respectively). The median time of regrowth was 86.1 (range, 23.2-236.0) months. Previous surgery was associated with tumor regrowth out of field (p = 0.033). Twelve patients underwent repeat GKRS, including regrowth in (n = 8) and out of field (n = 4). Tumor shrunk in seven patients (58.3%), remained stable in one (8.3%) and regrowth in four (33.3%) with a median repeat GKRS margin dose of 12 (range, 10.0-14.0) Gy. The actuarial tumor control rates were 100%, 90%, 90%, 68%, and 68% at 1, 3, 5, 10, and 15 years after repeat GKRS, respectively. CONCLUSIONS: Parasellar invasion and tumor margin dose (<12 Gy) were independent risk factors for tumor regrowth after GKRS. Repeat GKRS might be effective on tumor control for selected patients. For regrowth in field due to relatively insufficient radiation dose, repeat GKRS might offer satisfactory tumor control. For regrowth out of field, preventing regrowth out of field was the key management. Sufficient target coverage and close follow-up might be helpful.

Pituitary hemorrhage in pituitary adenomas treated with gamma knife radiosurgery: incidence, risk factors and prognosis.

Objective: The aims of this study were to investigate the incidence, risk factors and prognosis of pituitary hemorrhage in pituitary adenomas treated with gamma knife radiosurgery (GKRS). Methods and materials: Between December 1993 and December 2016, 751 consecutive pituitary adenoma patients treated with GKRS were retrospectively reviewed in a single center. There were 271 male (36.1%) and 480 female (63.9%) patients with a median age of 38.5 (range, 7.2-84.0) years. The number of nonfunctioning pituitary adenomas (NFPAs) and functioning pituitary adenomas were 369 (49.1%) and 382 (50.9%) respectively. The median follow-up time was 61.1 (range, 12.1-304.4) months. Results: In this study, 88 patients (11.7%) were diagnosed with pituitary hemorrhage before GKRS, 55 patients (7.3%) developed new or worsened pituitary hemorrhage after GKRS (excluding 3 patients with new or worsened pituitary hemorrhage due to tumor regrowth). The median time to new or worsened pituitary hemorrhage after GKRS was 18.9 (range 3.1-70.7) months. Overall, 128 patients (17.0%) were diagnosed with pituitary hemorrhage in the entire series. After adjustment with logistic regression, nonfunctioning pituitary adenomas (NFPAs) (odds ratio [OR]=2.121, 95% confidence interval [CI]=1.195-3.763, p=0.010) and suprasellar extension (OR=2.470, 95% CI=1.361-4.482, p=0.003) were associated with pituitary hemorrhage before GKRS. NFPA (OR=3.271, 95% CI=1.278-8.373, p=0.013) was associated with new or worsened pituitary hemorrhage after GKRS. Five patients received surgical resection for new or worsened pituitary hemorrhage were considered as GKRS treatment failure. Two patients with new hypopituitarism were considered to be owed to new or worsened pituitary hemorrhage after GKRS. Conclusions: New or worsened pituitary hemorrhage after GKRS was not an uncommon phenomenon. NFPA was an independent risk factor of new or worsened pituitary hemorrhage after GKRS. New or worsened pituitary hemorrhage after GKRS could lead to GKRS treatment failure. GKRS might be a precipitating factor of pituitary hemorrhage.

Initial Gamma Knife Radiosurgery for Large or Documented Growth Asymptomatic Meningiomas: Long-Term Results From a 27-Year Experience.

OBJECTIVE: The aims of this study were to investigate the long-term outcomes of initial Gamma Knife radiosurgery (GKRS) for large (≥20 mm) or documented growth asymptomatic meningiomas. DESIGN AND METHODS: This was a single-center retrospective study. Fifty-nine patients with large (≥20 mm) or documented growth asymptomatic meningiomas undergoing initial GKRS were enrolled. The median age was 56 (range, 27-83) years. The median time of follow-up was 66.8 (range, 24.6-245.6) months, and the median tumor margin dose was 13.0 Gy (range, 11.6-22.0 Gy). RESULTS: Tumors shrunk in 35 patients (59.3%) and remained stable in 23 (39.0%). One patient (1.7%) experienced radiological progression at 54 months after GKRS. The PFS was 100%, 97%, and 97% at 3, 5, and 10 years, respectively. Nine patients (15.3%) occurred new neurological symptoms or signs at a median time of 8.1 (range, 3.0-81.6) months. The symptom PFS was 90% and 78% at 5 and 10 years, respectively. Fifteen patients (25.4%) occurred peritumoral edema (PTE) at a median time of 7.2 (range, 2.0-81.6) months. One patient underwent surgical resection for severe PTE. In univariate and multivariate analysis, Only tumor size (≥25 mm) and maximum dose (≥34 Gy) were significantly associated with PTE [hazard ratio (HR)= 3.461, 95% confidence interval (CI)=1.157-10.356, p=0.026 and HR=3.067, 95% CI=1.068-8.809, P=0.037, respectively]. CONCLUSIONS: In this study, initial GKRS can provide a high tumor control rate as well as an acceptable rate of complications in large or documented growth asymptomatic meningiomas. GKRS may be an alternative initial treatment for asymptomatic meningiomas.

Gamma Knife Radiosurgery (GKRS) for Patients with Prolactinomas: Long-Term Results From a Single-Center Experience.

BACKGROUND The aim of this study was to review outcomes of gamma knife radiosurgery (GKRS) for prolactinoma and report our experience with it. MATERIAL AND METHODS We reviewed the patient database in our center and identified 24 patients with prolactinoma who underwent GKRS from 1993 to 2016.  Complete endocrine, clinical, and radiological data were available on these individuals before and after GKRS. RESULTS Data from 5 males and 19 females with a median age of 30.5 years (range, 18.1 to 51.1) were reviewed. The median follow-up was 109.3 months (range, 23.2-269.3). The median margin dose of GKRS was 15 Gy (range, 10.5 to 23.6). In total, prolactin (PRL) normalization after GKRS was achieved in 66.7% of patients. Endocrine remission (normal PRL levels after discontinuation of dopamine agonists) was achieved in 10 patients (41.7%), and endocrine control (normal PRL levels while taking dopamine agonists) was achieved in 6 patients (25.0%). All of the patients showed tumor control. New-onset hypopituitarism post-GKRS occurred in 4 patients (16.7%). No new visual dysfunction or cranial nerve dysfunction were observed after GKRS. CONCLUSIONS For treatment of prolactinomas, GKRS may provide relatively high rates of endocrine remission and tumor control, as well as a low rate of new-onset hypopituitarism. GKRS may be an effective and safe treatment for prolactinomas.

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

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.

Long-term results of Gamma Knife Radiosurgery for Postsurgical residual or recurrent nonfunctioning Pituitary Adenomas.

Introduction: The aim of this retrospective study was to analyze the long-term outcomes and factors associated with treatment failure of Gamma Knife radiosurgery (GKRS) for postsurgical residual or recurrent nonfunctioning pituitary adenomas (NFPAs). Design and Methods: A total of 148 cases of postsurgical residual or recurrent NFPA patients were enrolled in the study. There were 111 cases with residual tumor and 37 cases with recurrent tumor. The median age was 46.0 years (Range: 10.9-75.8 years). The median tumor volume at GKRS was 3.6 cm3 (Range: 0.3-74.5 cm3), and the median tumor margin dose was 14.0 Gy (Range: 9 - 20 Gy). Results: Tumor shrunk in 111 patients (75%), remained stable in 17 patients (11.5%), and progressed in 20 patients (13.5%) during a median of 64.5 months (Range: 14.5 - 236.0 months) of imaging follow-up. The progression-free survival rates were 99%, 91%, 88% and 74% at 1, 3, 5 and 10 years after GKRS, respectively. In a multivariate analysis, tumor margin dose (<13 Gy) was significantly associated with tumor progression (hazard ratio=3.526, 95% confidence interval=1.400-8.877, p=0.007). New hypopituitarism occurred in 22 out of 80 patients (27.5%), including hypogonadism (n=7), hypothyroidism (n=9), hypocortisolism (n=15) and growth hormone deficiency (n=1). In univariate and multivariate analysis, there were no factors significantly associated with new hypopituitarism. Six patients (4.1%) developed new or worsening visual dysfunction. Four patients (2.7%) developed new cranial neuropathy. Conclusion: In this study, GKRS can offer a high tumor control rate as well as a low rate of complications in postsurgical residual or recurrent NFPA patients.

A preliminary study to quantitatively evaluate the development of maturation degree for fetal lung based on transfer learning deep model from ultrasound images.

PURPOSE: The evaluation of fetal lung maturity is critical for clinical practice since the lung immaturity is an important cause of neonatal morbidity and mortality. For the evaluation of the development of fetal lung maturation degree, our study established a deep model from ultrasound images of four-cardiac-chamber view plane. METHODS: A two-stage transfer learning approach is proposed for the purpose of the study. A specific U-net structure is designed for the applied deep model. In the first stage, the model is to first learn the recognition of fetal lung region in the ultrasound images. It is hypothesized in our study that the development of fetal lung maturation degree is generally proportional to the gestational age. Then, in the second stage, the pretrained deep model is trained to accurately estimate the gestational age from the fetal lung region of ultrasound images. RESULTS: Totally 332 patients were included in our study, while the first 206 patients were used for training and the subsequent 126 patients were used for the independent testing. The testing results of the established deep model have the imprecision as 1.56 ± 2.17 weeks on the gestational age estimation. Its correlation coefficient with the ground truth of gestational age achieves 0.7624 (95% CI 0.6779 to 0.8270, P value < 0.00001). CONCLUSION: The hypothesis that the development of fetal lung maturation degree can be represented by the texture information from ultrasound images has been preliminarily validated. The fetal lung maturation degree can be considered as being represented by the deep model's output denoted by the estimated gestational age.

Initial Gamma Knife radiosurgery for nonfunctioning pituitary adenomas: results from a 26-year experience.

OBJECTIVE: The aim of this study was to evaluate the long-term outcomes of initial Gamma Knife radiosurgery (GKRS) for patients with nonfunctioning pituitary adenomas (NFPAs). DESIGN AND METHODS: This was a single-center retrospective study. Eighty-one patients with NFPAs undergoing initial GKRS were enrolled. The median age was 44.9 years (range, 7.2-75.5 years). The median tumor volume was 2.3 cm3 (range, 0.1-31.3 cm3), and the median tumor margin dose was 13.0 Gy (range, 8-22 Gy). RESULTS: Tumor shrunk in 63 patients (77.8%), remained stable in 9 (11.1%), treatment failure in 9 (11.1%) during a median follow-up of 67.1 months (range, 11.5-263.9 months). The tumor control rates were 100%, 99%, 95%, and 84%, at 1, 3, 5, and 10 years, respectively. In multivariate analysis, tumor volume (≥4 cm3) and margin dose (<12 Gy) were associated with treatment failure (hazard ratio (HR) = 7.093, 95% confidence interval (CI) = 1.098-45.083, p = 0.040, and HR = 9.643, 95% CI = 1.108-83.927, p = 0.040, respectively). New apoplexy occurred in seven patients (8.6%) after GKRS with a median time of 39.9 months (range, 11.9-166.8 months). In multivariate analysis, tumor volume (≥10 cm3) was a significant risk factor (HR = 10.642, 95% CI = 2.121-53.398, p = 0.004). New hypopituitarism occurred in 14 patients (17.3%). No factors were associated with new hypopituitarism. Four patients (4.9%) developed new or worsening visual dysfunction. No new cranial neuropathy was noted. CONCLUSIONS: In this study, initial GKRS can provide a high tumor control rate, as well as a low incidence rate of complications in NFPAs. GKRS may be an alternative initial treatment for selected NFPAs.

A hierarchical model for automated breast lesion detection from ultrasound 3D data.

Ultrasound imaging plays an important role in breast cancer screening for which early and accurate lesion detection is crucial for clinical practice. Many researches were performed on supporting the breast lesion detection based on ultrasound data. In the paper, a novel hierarchical model is proposed to automatically detect breast lesion from ultrasound 3D data. The model simultaneously considers the data information from low-level to high-level for the detection by processing with a joint probability. For each layer of the model, the corresponding algorithm is performed to denote the certain level image information. A dynamic programming approach is applied to efficiently obtain the optimal solution. With a preliminary dataset, the superior performance of the proposed model has been demonstrated for the automated detection of breast lesion with 0.375 false positive per case at 91.7% sensitivity.

Evaluation of fatty proportion in fatty liver using least squares method with constraints.

Backscatter and attenuation parameters are not easily measured in clinical applications due to tissue inhomogeneity in the region of interest (ROI). A least squares method(LSM) that fits the echo signal power spectra from a ROI to a 3-parameter tissue model was used to get attenuation coefficient imaging in fatty liver. Since fat's attenuation value is higher than normal liver parenchyma, a reasonable threshold was chosen to evaluate the fatty proportion in fatty liver. Experimental results using clinical data of fatty liver illustrate that the least squares method can get accurate attenuation estimates. It is proved that the attenuation values have a positive correlation with the fatty proportion, which can be used to evaluate the syndrome of fatty liver.

Ultrasound lesion segmentation using clinical knowledge-driven constrained level set.

Ultrasound lesion segmentation is an important and challenging task. Comparing with other methods, region-based level set has many advantages, but still requires considerable improvement to deal with the characteristic of lesions in the ultrasound modality such as shadowing, speckle and heterogeneity. In the clinical workflow, the physician would usually denote long and short axes of a lesion for measurement purpose yielding four markers in an image. Inspired by this workflow, a constrained level set method is proposed to fully utilize these four markers as prior knowledge and global constraint for the segmentation. First, the markers are detected using template-matching algorithm and B-Spline is applied to fit four markers as the initial contour. Then four-marker constrained energy is added to the region-based local level set to make sure that the contour evolves without deviation from the four markers. Finally the algorithm is implemented in a multi-resolution scheme to achieve sufficient computational efficiency. The performance of the proposed segmentation algorithm was evaluated by comparing our results with manually segmented boundaries on 308 ultrasound images with breast lesions. The proposed method achieves Dice similarity coefficient 89.49 ± 4.76% and could be run in real-time.