Role of Procalcitonin as a Biomarker in Early Identification of Adverse Events Following Esophageal Atresia Surgery.

Introduction: Surgical complication following esophageal atresia repair is one of the several factors known to influence the final outcomes. Early identification of such complications may help in timely institution of therapeutic measures and translate into improved prognosis. Objective: The objective of this study was to evaluate the role of procalcitonin in early prediction of the adverse events after surgery in patients of esophageal atresia and the temporal relationship with clinical manifestations and other inflammatory biomarkers such as C-reactive protein (CRP). Materials and Methods: This was a prospective study on consecutive patients of esophageal atresia (n = 23). Serum procalcitonin and CRP levels were assessed at baseline (prior to surgery) and on postoperative days (POD) 1, 3, 5, 7, and 14. The trends in the biomarker values and temporal relationships of deviation in trend with the clinical and conventional laboratory parameters and patient outcomes were analyzed. Results: Baseline serum procalcitonin was elevated (n = 23; 1.7 ng/ml: min: 0.07 ng/ml-max: 24.36 ng/ml) in 18/23 (78.3%) patients. Procalcitonin nearly doubled on POD-1 (n = 22; 3.28 ng/ml: min: 0.64 ng/ml-max: 16.51 ng/ml) followed by a gradual decline. CRP was also elevated on POD-1 (three times the baseline) and depicted a delayed peak at POD-3. POD-1 procalcitonin and CRP levels correlated with survival. POD-1 procalcitonin cutoff at 3.28 ng/ml predicted mortality with a sensitivity and specificity of 100% and 57.9% (P = 0.05). Serum procalcitonin and CRP were higher for patients who sustained complications, so was the time required for hemodynamic stabilization. Procalcitonin (baseline and POD-5) and CRP (POD-3 and POD-5) values correlated with the clinical course after surgery. Baseline procalcitonin cutoff at 2.91 ng/ml predicted the possibility of a major complication with a sensitivity of 71.4% and a specificity of 93.3%. POD-5 procalcitonin cutoff at 1.38 ng/ml predicted the possibility of a major complication with a sensitivity of 83.3% and a specificity of 93.3%. Patients who sustained major complications depicted a change in serum procalcitonin trend 24-48 h ahead of clinical manifestation of an adverse event. Conclusions: Procalcitonin is a good indicator to identify the adverse events in neonates after surgery for esophageal atresia. The procalcitonin levels in patients who sustained a major complication depicted a reversal in trend 24-48 h of clinical manifestation. POD-1 procalcitonin correlated with survival while the baseline and POD-5 serum procalcitonin predicted the clinical course.

Renal Epithelioid Angiomyolipoma in Children.

Renal angiomyolipoma is a rare cause of renal tumor in children. Most are associated with tuberous sclerosis, and the classic type is observed more commonly. Epithelioid angiomyolipoma is even rarer with only limited case reports and series published in literature, most of which are of adult patients. We describe a 12-year-old boy, a diagnosed patient of tuberous sclerosis, who presented with pain in the left flank. On evaluation, it was found to have a left renal mass with the clinical picture suggestive of renal cell carcinoma. Partial nephrectomy was performed and histopathology revealed epithelioid angiomyolipoma. The child was asymptomatic at follow-up after 3 months. Only a few such cases in children are found in literature, which are discussed alongside. Differential diagnosis of this rare tumor must be kept in mind in a renal tumor as surgery is generally curative in this possibly malignant tumor. Metastasis confers a poor prognosis. Chemotherapy is generally not effective, although various regimens have been tried. Tumor recurrence must be kept in mind and a follow-up after apparent complete remission is of paramount importance.

Surgical outcomes in oral cancer involving the central arch of the mandible in elderly patients: An institutional experience.

BACKGROUNDS: Oral squamous cell cancer (SCC) is one of the most common cancers. The most common age of presentation is fifth to sixth decade. Management of this disease is dictated by stage, age, and related comorbidities. Elderly patients have their own set of limitations as far as their management is concerned. Carcinoma involving central mandibular arch is a challenging disease for surgeons, especially in the elderly. This article describes our experience with the surgical treatment of oral cancer involving the central arch of the mandible in elderly patients. METHODS: Forty elderly (≥60 years) patients with histologically proven SCC of the oral cavity in which disease was involving the central arch of the mandible, were included in our study. Demographic, clinical, and treatment-related factors were recorded. The outcome was assessed in terms of postoperative complications, recurrence, and patient survival. RESULTS: The median age of the patients was 63 years. The male:female ratio was 7:3. A history of oral tobacco use was present in 95% of patients. The most common site of disease was lower alveolus (80%) followed by carcinoma of the lower lip (20%). Majority of our patients (77.3% [30]) were having Stage IV disease. Mandibulectomy was either segmental (62.5%) or marginal (37.5%). Bilateral neck dissection (37/40, 92.5%) was done in most patients. Among all patients, 62.5% (25) received adjuvant radiotherapy. The local recurrence rate after a median follow-up of 30 months was 15% (6). Two-year disease-free survival and overall survival were 89% and 90%, respectively. CONCLUSION: Central arch of the mandible is a difficult disease to treat. It needs a complex and lengthy reconstructive procedure. Comorbidities such as extreme age, diabetes, and pulmonary and cardiac illnesses make it more challenging to manage. With the proper evaluation of comorbidities and avoiding long, cumbersome procedures, we can provide patients a fairly good chance of survival.

Distributed Newton Methods for Deep Neural Networks.

Deep learning involves a difficult nonconvex optimization problem with a large number of weights between any two adjacent layers of a deep structure. To handle large data sets or complicated networks, distributed training is needed, but the calculation of function, gradient, and Hessian is expensive. In particular, the communication and the synchronization cost may become a bottleneck. In this letter, we focus on situations where the model is distributedly stored and propose a novel distributed Newton method for training deep neural networks. By variable and feature-wise data partitions and some careful designs, we are able to explicitly use the Jacobian matrix for matrix-vector products in the Newton method. Some techniques are incorporated to reduce the running time as well as memory consumption. First, to reduce the communication cost, we propose a diagonalization method such that an approximate Newton direction can be obtained without communication between machines. Second, we consider subsampled Gauss-Newton matrices for reducing the running time as well as the communication cost. Third, to reduce the synchronization cost, we terminate the process of finding an approximate Newton direction even though some nodes have not finished their tasks. Details of some implementation issues in distributed environments are thoroughly investigated. Experiments demonstrate that the proposed method is effective for the distributed training of deep neural networks. Compared with stochastic gradient methods, it is more robust and may give better test accuracy.

A theory of frequency domain invariants: spherical harmonic identities for BRDF/lighting transfer and image consistency.

This paper develops a theory of frequency domain invariants in computer vision. We derive novel identities using spherical harmonics, which are the angular frequency domain analog to common spatial domain invariants such as reflectance ratios. These invariants are derived from the spherical harmonic convolution framework for reflection from a curved surface. Our identities apply in a number of canonical cases, including single and multiple images of objects under the same and different lighting conditions. One important case we consider is two different glossy objects in two different lighting environments. For this case, we derive a novel identity, independent of the specific lighting configurations or BRDFs, that allows us to directly estimate the fourth image if the other three are available. The identity can also be used as an invariant to detecttampering in the images. While this paper is primarily theoretical, it has the potential to lay the mathematical foundations for two important practical applications. First, we can develop more general algorithms for inverse rendering problems, which can directly relight and change material properties by transferring the BRDF or lighting from another object or illumination. Second, we can check the consistency of an image, to detect tampering or image splicing.