Use of natural language processing to extract and classify papillary thyroid cancer features from surgical pathology reports.

BACKGROUND: We aim to use Natural Language Processing (NLP) to automate the extraction and classification of thyroid cancer risk factors from pathology reports. METHODS: We analyzed 1,410 surgical pathology reports from adult papillary thyroid cancer patients from 2010 to 2019. Structured and non-structured reports were used to create a consensus-based ground truth dictionary and categorized them into modified recurrence risk levels. Non-structured reports were narrative, while structured reports followed standardized formats. We developed ThyroPath, a rule-based NLP pipeline, to extract and classify thyroid cancer features into risk categories. Training involved 225 reports (150 structured, 75 unstructured), with testing on 170 reports (120 structured, 50 unstructured) for evaluation. The pipeline's performance was assessed using both strict and lenient criteria for accuracy, precision, recall, and F1-score, a metric that combines precision and recall evaluation. RESULTS: In extraction tasks, ThyroPath achieved overall strict F-1 scores of 93% for structured reports and 90% for unstructured reports, covering 18 thyroid cancer pathology features. In classification tasks, ThyroPath-extracted information demonstrated an overall accuracy of 93% in categorizing reports based on their corresponding guideline-based risk of recurrence: 76.9% for high-risk, 86.8% for intermediate risk, and 100% for both low and very low-risk cases. However, ThyroPath achieved 100% accuracy across all risk categories with human extracted pathology information. CONCLUSIONS: ThyroPath shows promise in automating the extraction and risk recurrence classification of thyroid pathology reports at large scale. It offers a solution to laborious manual reviews and advancing virtual registries. However, it requires further validation before implementation.

Effects of Optimal Temperature Control in Body Contouring Surgery: A Non-Randomized Controlled Clinical Trial.

BACKGROUND: Perioperative hypothermia in plastic surgery has underestimated risks, including increased risk of infection, cardiac events, blood loss, prolonged recovery time, increased nausea, pain, and opioid usage. Inadequate preventive measures can result in up to 4 hours of normothermia restoration. OBJECTIVES: Compare the impact of different strategies for normothermia during plastic surgery procedures and its relationship with clinical outcomes. METHODS: A non-randomized clinical trial was conducted in a single center in Bogota, Colombia. We enrolled adult patients undergoing body contouring surgery and divided them into four intervention groups with different measures to control body temperature. Univariate and Bivariate analyses were performed comparing several clinical symptoms to evaluate outcomes. RESULTS: A total of 197 patients were analyzed. Most of them were women (84,3%). Mean age was 38.6 years, and a median procedure duration of 260 minutes. Demographic and clinical characteristics did not exhibit significant differences between the groups. However, there were notable variations in temperature measurements at crucial moments during the surgical procedure among the groups, attributed to the implementation of distinct thermal protective strategies. Group comparisons showed a relationship between hypothermia with increased nausea, vomiting, shivering, pain, and additional analgesia requirements. CONCLUSIONS: Incorporation of active thermal protective measures, such as Blanketrol or HotDog, during body contouring procedures, markedly diminishes the risk of hypothermia and enhances overall clinical outcomes. Implementing these active measures to maintain the patient in a state of normothermia not only improves operating room efficiency but also leads to a reduction in recovery room duration.

Mesh-like External Suture Splint: A Way to Improve Results and Reduce Postoperative Complications after High-definition Liposculpture.

High-definition liposculpture with variable degrees of muscular definition allows both patients and surgeons to broaden the procedure to fit almost any individual preferences; however, skin laxity and those secondary cases still represent a big challenge for the surgeon to achieve optimal outcomes. We are reporting our experience with a new suture-splint procedure to reduce edema, avoid skin laxity, and increase muscular definition after liposculpture. We enrolled patients undergoing HDL with any risk factor that could compromise aesthetic outcomes. Tumescent technique with VASER fat emulsification and power-assisted liposuction were used to complete HDL. We performed a continuous suture over the midline and other areas of muscular definition right before patients awakened from anesthesia. Real-time ultrasound imaging was performed to ensure the safety of the procedure. A total of 52 consecutive patients were treated with MesHD in our study. Only two patients voluntarily dropped off from postoperative analysis due to geographical limitations. Postoperative edema was referred by only one patient after surgery. No major complications were reported. Three patients complained of stabbing pain over one of the stitches, which resolved completely after suture removal. All patients were satisfied with postoperative results. Mesh suture-splint at body areas with high risk of adverse effects after body contouring surgery may be an effective procedure to improve recovery by decreasing edema formation, enhancingmuscular definition, decreasing risk of fibrosis, and avoiding skin laxity postoperative appearance.

Significance of Pathologic Response in Patients With Early and Locally Advanced Breast Cancer Treated With Neoadjuvant Chemotherapy in a Middle-Income Country. A Real-World Historical Cohort.

PURPOSE: Breast cancer (BC) is the most frequent neoplasm in women in Colombia and is associated with a higher mortality rate than in other countries and regions. Neoadjuvant chemotherapy (NACT) has become a standard treatment in locally advanced BC and provides an opportunity to improve clinical outcomes in BC. This study aims to describe characteristics, treatment patterns, and outcomes after NACT in a cohort of Colombian patients with BC. METHODS: We performed a retrospective cohort study. We included adult patients with BC treated with NACT. Clinical charts were retrospectively reviewed. Descriptive statistics and time to event for overall survival analyses were performed. Recursive partitioning was performed for survival curves to assess the complex relationship between survival times and other variables. RESULTS: Three hundred and fourteen patients were included for analysis. The pathologic complete response after neoadjuvant chemotherapy (ypCR) rate was 34.4%, with a higher ypCR in triple-negative BC (TNBC; 46.9%) and human epidermal growth factor receptor 2-positive BC (72.7%). Those who did not achieve ypCR had a higher percentage of death and relapse. The median follow-up was 4.9 years, with an 88.2% 5-year overall survival (OS). CONCLUSION: A total of 62.6% of the total patients identified were not treated with NACT, indicating a low utilization. Our global ypCR rate was higher when compared with similar studies in Colombia, likely because of differences in the NACT treatment regimens. ypCR was only associated with OS in the TNBC subgroup, emphasizing the importance of pursuing ypCR in these patients. We consider the use of NACT a valuable opportunity to implement innovative treatment approaches that improve outcomes in Colombian patients with BC.

A Systematic Review of Natural Language Processing Methods and Applications in Thyroidology.

This study aimed to review the application of natural language processing (NLP) in thyroid-related conditions and to summarize current challenges and potential future directions. We performed a systematic search of databases for studies describing NLP applications in thyroid conditions published in English between January 1, 2012 and November 4, 2022. In addition, we used a snowballing technique to identify studies missed in the initial search or published after our search timeline until April 1, 2023. For included studies, we extracted the NLP method (eg, rule-based, machine learning, deep learning, or hybrid), NLP application (eg, identification, classification, and automation), thyroid condition (eg, thyroid cancer, thyroid nodule, and functional or autoimmune disease), data source (eg, electronic health records, health forums, medical literature databases, or genomic databases), performance metrics, and stages of development. We identified 24 eligible NLP studies focusing on thyroid-related conditions. Deep learning-based methods were the most common (38%), followed by rule-based (21%), and traditional machine learning (21%) methods. Thyroid nodules (54%) and thyroid cancer (29%) were the primary conditions under investigation. Electronic health records were the dominant data source (17/24, 71%), with imaging reports being the most frequently used (15/17, 88%). There is increasing interest in NLP applications for thyroid-related studies, mostly addressing thyroid nodules and using deep learning-based methodologies with limited external validation. However, none of the reviewed NLP applications have reached clinical practice. Several limitations, including inconsistent clinical documentation and model portability, need to be addressed to promote the evaluation and implementation of NLP applications to support patient care in thyroidology.

Thyroid Ultrasound Appropriateness Identification Through Natural Language Processing of Electronic Health Records.

Objective: To address thyroid cancer overdiagnosis, we aim to develop a natural language processing (NLP) algorithm to determine the appropriateness of thyroid ultrasounds (TUS). Patients and Methods: Between 2017 and 2021, we identified 18,000 TUS patients at Mayo Clinic and selected 628 for chart review to create a ground truth dataset based on consensus. We developed a rule-based NLP pipeline to identify TUS as appropriate TUS (aTUS) or inappropriate TUS (iTUS) using patients' clinical notes and additional meta information. In addition, we designed an abbreviated NLP pipeline (aNLP) solely focusing on labels from TUS order requisitions to facilitate deployment at other health care systems. Our dataset was split into a training set of 468 (75%) and a test set of 160 (25%), using the former for rule development and the latter for performance evaluation. Results: There were 449 (95.9%) patients identified as aTUS and 19 (4.06%) as iTUS in the training set; there are 155 (96.88%) patients identified as aTUS and 5 (3.12%) were iTUS in the test set. In the training set, the pipeline achieved a sensitivity of 0.99, specificity of 0.95, and positive predictive value of 1.0 for detecting aTUS. The testing cohort revealed a sensitivity of 0.96, specificity of 0.80, and positive predictive value of 0.99. Similar performance metrics were observed in the aNLP pipeline. Conclusion: The NLP models can accurately identify the appropriateness of a thyroid ultrasound from clinical documentation and order requisition information, a critical initial step toward evaluating the drivers and outcomes of TUS use and subsequent thyroid cancer overdiagnosis.

The cancer-related transcription factor Runx2 modulates cell proliferation in human osteosarcoma cell lines.

Runx2 regulates osteogenic differentiation and bone formation, but also suppresses pre-osteoblast proliferation by affecting cell cycle progression in the G(1) phase. The growth suppressive potential of Runx2 is normally inactivated in part by protein destabilization, which permits cell cycle progression beyond the G(1)/S phase transition, and Runx2 is again up-regulated after mitosis. Runx2 expression also correlates with metastasis and poor chemotherapy response in osteosarcoma. Here we show that six human osteosarcoma cell lines (SaOS, MG63, U2OS, HOS, G292, and 143B) have different growth rates, which is consistent with differences in the lengths of the cell cycle. Runx2 protein levels are cell cycle-regulated with respect to the G(1)/S phase transition in U2OS, HOS, G292, and 143B cells. In contrast, Runx2 protein levels are constitutively expressed during the cell cycle in SaOS and MG63 cells. Forced expression of Runx2 suppresses growth in all cell lines indicating that accumulation of Runx2 in excess of its pre-established levels in a given cell type triggers one or more anti-proliferative pathways in osteosarcoma cells. Thus, regulatory mechanisms controlling Runx2 expression in osteosarcoma cells must balance Runx2 protein levels to promote its putative oncogenic functions, while avoiding suppression of bone tumor growth.

Evaluation of factors leading to poor outcomes for pediatric acute lymphoblastic leukemia in Mexico: a multi-institutional report of 2,116 patients.

Background and aims: Pediatric acute lymphoblastic leukemia (ALL) survival rates in low- and middle-income countries are lower due to deficiencies in multilevel factors, including access to timely diagnosis, risk-stratified therapy, and comprehensive supportive care. This retrospective study aimed to analyze outcomes for pediatric ALL at 16 centers in Mexico. Methods: Patients <18 years of age with newly diagnosed B- and T-cell ALL treated between January 2011 and December 2019 were included. Clinical and biological characteristics and their association with outcomes were examined. Results: Overall, 2,116 patients with a median age of 6.3 years were included. B-cell immunophenotype was identified in 1,889 (89.3%) patients. The median white blood cells at diagnosis were 11.2.5 × 103/mm3. CNS-1 status was reported in 1,810 (85.5%), CNS-2 in 67 (3.2%), and CNS-3 in 61 (2.9%). A total of 1,488 patients (70.4%) were classified as high-risk at diagnosis. However, in 52.5% (991/1,889) of patients with B-cell ALL, the reported risk group did not match the calculated risk group allocation based on National Cancer Institute (NCI) criteria. Fluorescence in situ hybridization (FISH) and PCR tests were performed for 407 (19.2%) and 736 (34.8%) patients, respectively. Minimal residual disease (MRD) during induction was performed in 1,158 patients (54.7%). The median follow-up was 3.7 years. During induction, 191 patients died (9.1%), and 45 patients (2.1%) experienced induction failure. A total of 365 deaths (17.3%) occurred, including 174 deaths after remission. Six percent (176) of patients abandoned treatment. The 5-year event-free survival (EFS) was 58.9% ± 1.7% for B-cell ALL and 47.4% ± 5.9% for T-cell ALL, while the 5-year overall survival (OS) was 67.5% ± 1.6% for B-cell ALL and 54.3% ± 0.6% for T-cell ALL. The 5-year cumulative incidence of central nervous system (CNS) relapse was 5.5% ± 0.6%. For the whole cohort, significantly higher outcomes were seen for patients aged 1-10 years, with DNA index >0.9, with hyperdiploid ALL, and without substantial treatment modifications. In multivariable analyses, age and Day 15 MRD continued to have a significant effect on EFS. Conclusion: Outcomes in this multi-institutional cohort describe poor outcomes, influenced by incomplete and inconsistent risk stratification, early toxic death, high on-treatment mortality, and high CNS relapse rate. Adopting comprehensive risk-stratification strategies, evidence-informed de-intensification for favorable-risk patients and optimized supportive care could improve outcomes.

Securing access to a comprehensive diagnostic panel for children with suspected acute lymphoblastic leukemia: Results from the Mexico in Alliance with St. Jude "Bridge Project".

Background: The "Bridge Project" is a Mexico in Alliance with St. Jude (MAS) initiative developed in 2019 to improve access, accuracy, and timeliness of specialized diagnostic studies for patients with suspected acute lymphoblastic leukemia (ALL). The project strategy relies on service centralization to improve service delivery, biological characterization, risk-group classification, and support proper treatment allocation. Methods: This is an ongoing prospective multisite intersectoral quality improvement (QI) project available to all patients 0-18 years of age presenting with suspected ALL to the 14 actively participating institutions in 12 Mexican states. Institutions send specimens to one centralized laboratory. From a clinical standpoint, the project secures access to a consensus-derived comprehensive diagnostic panel. From a service delivery standpoint, we assess equity, timeliness, effectiveness, and patient-centeredness. From an implementation science standpoint, we document feasibility, utility, and appropriateness of the diagnostic panel and centralized approach. This analysis spans from July 2019 to June 2023. Results: 612 patients have accessed the project. The median age was 6 years (IQR 3-11), and 53% were males. 94% of the specimens arrived within 48 hours, which documents the feasibility of the centralized model, and 100% of the patients received precise and timely diagnostic results, which documents the effectiveness of the approach. Of 505 (82.5%) patients with confirmed ALL, 463/505 (91.6%) had B-cell ALL, and 42/505 (8.3%) had T-cell ALL. High-hyperdiploidy was detected by DNA index in 36.6% and hypodiploidy in 1.6%. 76.6% of the patients had conclusive karyotype results. FISH studies showed t(12;21) in 15%, iAMP21 in 8.5%, t(1;19) in 7.5%, t(4;11) in 4.2%, t(9;22) in 3.2%, del(9)(p21) in 1.8%, and TRA/D (14)(q11.2) rearrangement in 2.4%. Among B-cell ALL patients, 344/403 (85.1%) had Day 15 MRD<1% and 261/305 (85.6%) Day 84 MRD<0.01. For T-cell ALL patients 20/28 (71.4%) had Day 29 MRD<0.01% and 19/22 (86.4%) Day 84 MRD<0.01%. Conclusions: By securing access to a standardized consensus-derived diagnostic panel, the Bridge Project has allowed better characterization of childhood ALL in Mexico while producing unprecedented service improvements and documenting key implementation outcomes. We are using these results to inform iterative changes to the diagnostic panel and an associated treatment guideline (MAS-ALL18).

The Influence of Angle Alpha, Angle Kappa, and Optical Aberrations on Visual Outcomes after the Implantation of a High-Addition Trifocal IOL.

The aim of our investigation was to examine the possible correlations between optical aberrations, angle kappa, angle alpha, and visual outcomes following cataract surgery. In total, 56 eyes of 28 patients were implanted with the Liberty 677MY trifocal intraocular lens (IOL). Pre- and postoperative higher-order aberrations, coma, astigmatism, angle alpha, and angle kappa were registered, along with uncorrected and corrected visual acuities at multiple distances. Visual acuity and contrast sensitivity defocus curves were plotted, and the areas under the curve were calculated 1 and 3 months postoperatively. Excellent visual outcomes were found at all distances. Patients reported low levels of dysphotopsia, and 96.4% of patients achieved complete spectacle independence. While angle kappa significantly decreased during cataract surgery (p = 0.0007), angle alpha remained unchanged (p = 0.5158). Angle alpha correlated with postoperative HOAs and had a negative impact on near vision (p = 0.0543). Preoperative corneal HOA and coma had a strong adverse effect on future intermediate and near vision. Residual astigmatism significantly affected postoperative intermediate vision (p = 0.0091). Our results suggest that angle kappa is not an optimal predictive factor for future visual outcomes, while angle alpha and the preoperative screening of optical aberrations might help patient selection prior to multifocal IOL implantation.

Cell-Free Glycoengineering of the Recombinant SARS-CoV-2 Spike Glycoprotein.

The baculovirus-insect cell expression system is readily utilized to produce viral glycoproteins for research as well as for subunit vaccines and vaccine candidates, for instance against SARS-CoV-2 infections. However, the glycoforms of recombinant proteins derived from this expression system are inherently different from mammalian cell-derived glycoforms with mainly complex-type N-glycans attached, and the impact of these differences in protein glycosylation on the immunogenicity is severely under investigated. This applies also to the SARS-CoV-2 spike glycoprotein, which is the antigen target of all licensed vaccines and vaccine candidates including virus like particles and subunit vaccines that are variants of the spike protein. Here, we expressed the transmembrane-deleted human ß-1,2 N-acetlyglucosamintransferases I and II (MGAT1ΔTM and MGAT2ΔTM) and the ß-1,4-galactosyltransferase (GalTΔTM) in E. coli to in-vitro remodel the N-glycans of a recombinant SARS-CoV-2 spike glycoprotein derived from insect cells. In a cell-free sequential one-pot reaction, fucosylated and afucosylated paucimannose-type N-glycans were converted to complex-type galactosylated N-glycans. In the future, this in-vitro glycoengineering approach can be used to efficiently generate a wide range of N-glycans on antigens considered as vaccine candidates for animal trials and preclinical testing to better characterize the impact of N-glycosylation on immunity and to improve the efficacy of protein subunit vaccines.

Impaired cell cycle regulation of the osteoblast-related heterodimeric transcription factor Runx2-Cbfbeta in osteosarcoma cells.

Bone formation and osteoblast differentiation require the functional expression of the Runx2/Cbfbeta heterodimeric transcription factor complex. Runx2 is also a suppressor of proliferation in osteoblasts by attenuating cell cycle progression in G(1). Runx2 levels are modulated during the cell cycle, which are maximal in G(1) and minimal beyond the G(1)/S phase transition (S, G(2), and M phases). It is not known whether Cbfbeta gene expression is cell cycle controlled in preosteoblasts nor how Runx2 or Cbfbeta are regulated during the cell cycle in bone cancer cells. We investigated Runx2 and Cbfbeta gene expression during cell cycle progression in MC3T3-E1 osteoblasts, as well as ROS17/2.8 and SaOS-2 osteosarcoma cells. Runx2 protein levels are reduced as expected in MC3T3-E1 cells arrested in late G(1) (by mimosine) or M phase (by nocodazole), but not in cell cycle arrested osteosarcoma cells. Cbfbeta protein levels are cell cycle independent in both osteoblasts and osteosarcoma cells. In synchronized MC3T3-E1 osteoblasts progressing from late G1 or mitosis, Runx2 levels but not Cbfbeta levels are cell cycle regulated. However, both factors are constitutively elevated throughout the cell cycle in osteosarcoma cells. Proteasome inhibition by MG132 stabilizes Runx2 protein levels in late G(1) and S in MC3T3-E1 cells, but not in ROS17/2.8 and SaOS-2 osteosarcoma cells. Thus, proteasomal degradation of Runx2 is deregulated in osteosarcoma cells. We propose that cell cycle control of Runx2 gene expression is impaired in osteosarcomas and that this deregulation may contribute to the pathogenesis of osteosarcoma.