In spite of its relatively straightforward diagnosis, which includes clinical/subclinical hyperthyroidism with or without goiter, increased free thyroxine and nonsuppressed TSH levels, and pituitary mass, the diagnosis of TSH-secreting and cosecreting adenomas was frequently unrecognized and thus much delayed. Serum alpha-subunit levels were high in nearly all patients with TSH-secreting adenomas and useful in excluding other conditions in the differential diagnosis. Proper indication and interpretation of simple laboratory tests should be emphasized in medical education to improve diagnostic accuracy.

No abstract

Context Artificial intelligence (AI), in particular machine learning (ML), may be used to deeply analyze biomarkers of response to first-generation somatostatin receptor ligands (fg-SRLs) in the treatment of acromegaly. Aim To develop a prediction model of therapeutic response of acromegaly to fg-SRL. Methods Patients with acromegaly not cured by primary surgical treatment and who had adjuvant therapy with fg-SRL for at least 6 months after surgery were included. Patients were considered controlled if they presented GH < 1.0 ng/mL and normal age-adjusted IGF-I levels. Six AI models were evaluated: logistic regression, k-nearest neighbor classifier, support vector machine, gradient-boosted classifier, random forest and multilayer perceptron. The features included in the analysis were age at diagnosis, sex, GH and IGF-I levels at diagnosis and at pretreatment, somatostatin receptor subtype 2 and 5 (SST2 and SST5) protein expression and cytokeratin granulation pattern (GP). Results A total of 153 patients were analyzed. Controlled patients were older (p = 0.002), had lower GH at diagnosis (p = 0.01), had lower pretreatment GH and IGF-I (p < 0.001), and more frequently harbored tumors that were densely granulated (p = 0.014) or highly expressed SST2 (p < 0.001).The model that performed best was the support vector machine with the features SST2, SST5, GP, sex, age, and pretreatment GH and IGF-I levels. It had an accuracy of 86.3%, positive predictive value of 83.3% and negative predictive value of 87.5%. Conclusion We developed a ML-based prediction model with high accuracy that has the potential to improve medical management of acromegaly, optimize biochemical control, decrease long-term morbidities and mortality and reduce health services costs.

Background: Long-term remission of acromegaly after somatostatin analog withdrawal has been reported in 18-42% of patients in studies with a relatively small number of patients using different inclusion and remission criteria. The objectives of this study were to establish the probability and predictive factors for short- and long-term remission [normal IGF-1 for age/sex: IGF-1 ≤1.00 × upper limit of normal (ULN)] after octreotide long-acting release (LAR) withdrawal in a larger population of well-controlled patients with acromegaly (normal mean IGF-1 in the last 24 months). Methods: This is a prospective multicenter study in which 58 well-controlled patients with acromegaly receiving only octreotide LAR as a primary or postsurgical treatment were included in 14 university centers in Brazil. All patients had been on stable doses and dose intervals of octreotide LAR in the last year, and none had been submitted to radiotherapy. The main outcome measure was serum IGF-1 after 8 weeks (short-term) and 60 weeks (long-term) of octreotide LAR withdrawal. Results: Seventeen of 58 patients (29%) were in remission in the short term, and only 4 patients achieved long-term remission after treatment withdrawal. The Kaplan-Meier estimated remission probability at 60 weeks was 7% and decreased to 5% at 72 weeks. The short-term remission rate was significantly higher (44%; p = 0.017) in patients with pretreatment IGF-1 <2.4 × ULN. No other predictive factor for short- or long-term remission was found. Conclusion: Our results show that long-term remission of acromegaly after octreotide LAR withdrawal was an uncommon and frequently unsustainable event and do not support the recommendation of a systematic withdrawal of treatment in controlled patients.

-Context -Carcinoembryonic antigen (CEA) can be detected in colorectal tumor tissue but its role in the survival of patients remains controversial. Objective -To characterize the expression of tissue CEA using immunohistochemical staining in colorectal tumors and to analyze the relationship between this finding and preoperative plasmatic level of CEA, morphologic features and survival of patients operated with curative intent for colorectal carcinoma. Method -Forty-seven patients were included in the study: 18 (38.3%) males and 29 (61.7%) females, with a mean age of 67.8 ± 9.7 years (37 to 84 years). Immediately before laparotomy, pre-operative serum levels of CEA were obtained where normal levels were considered ≤2.5 ng/mL for non-smokers, and ≤5.0 ng/mL for smokers. CEA immunohistochemical studies were carried out using anti-human CEA monoclonal mouse antibody. The expression of immunostaining for each neoplasia was classified according to the pattern of CEA tissular distribution into apical or cytoplasmic. The variables considered for the statistical analysis were plasmatic preoperative CEA level, location of the lesion within the large intestine, lesion diameter, lymph node involvement, Duke's classification, vein invasion, grade of cellular differentiation, survival and pattern of CEA tissular distribution. The statistical models utilized were Spearman's correlation and the Mann-Whitney, Kruskal-Wallis and Student t tests. Patients' survival was analyzed using the KaplanMeier method. Results -The mean preoperative CEA value was 15.4 ± 5.5 ng/mL (0.2 to 92.1 ng/mL). The neoplasm was located in the colon in 29 (61.7%) and in the rectum in 18 (38.3%) patients. Eight (17.0%) patients were classified as Duke's stage A, 22 (46.8%) as stage B and 17 (36.2%) as stage C. On immunohistochemical studies, the pattern of CEA tissular distribution was apical in 33 (70.2%) patients and cytoplasmic in 14 (29.8%) patients. Patients with apical patterns presented a mean sera CEA level of 15.5 ± 6.5 ng/mL while those with cytoplasmic pattern attained a mean sera CEA level of 15.1 ± 7.3 ng/mL, with no significant difference between these values (P = 0.35). Apical distribution of CEA occurred in 6 (12.8%) Duke A, 18 (38.2%) Duke B and 9 (12.2%) Duke C patients, while cytoplasmic CEA tissular distribution was observed in 2 (4.2%) Duke A, 3 (6.4%) Duke B and 9 (19.1%) Duke C patients. Patients with Duke B neoplasms presented significantly more apical CEA tissular distribution patterns (P = 0.049) than subjects with cytoplasmic CEA tissular patterns. The apical CEA tissular distribution pattern in neoplasms was significantly more frequent in neoplasms with no lymph node compromise compared to the cytoplasmic pattern (P = 0.50). However, no significant differences were seen between apical and cytoplasmic CEA tissular distribution patterns in terms of colon or rectal site (P = 0.21), lesion diameter across greatest axis (P = 0.19), vein invasion (P = 0.13) or degree of cellular differentiation (P = 0.19). Of the 47 patients...