INTRODUÇÃO: O câncer do reto médio e distal ainda é assunto bastante controverso, especialmente no que se refere ao estadiamento locorregional e opções terapêuticas. Busca-se um método sensível e específico para a avaliação da profundidade de invasão da parede retal e o envolvimento linfonodal. O adequado estadiamento da neoplasia do reto extraperitoneal é de suma importância no manejo terapêutico e prognóstico do paciente. Diversos métodos têm sido descritos para a avaliação da disseminação locorregional das neoplasias do reto, que variam desde o toque retal até a ressonância magnética da pelve e a ultrassonografia endorretal bi e tridimensional. OBJETIVO: Correlacionar os achados da profundidade de invasão tumoral na parede retal (T), comprometimento linfonodal (N), extensão e porcentagem de acometimento da lesão à ultrassonografia endorretal tridimensional (USER-3D) com o exame anatomopatológico (AP) de pacientes portadores de neoplasia de reto extraperitoneal submetidos a procedimento cirúrgico após o diagnóstico e estadiamento clínico prévio. MÉTODO: Estudo prospectivo foi realizado com pacientes portadores de neoplasia de reto médio e distal seguidos no Instituto Central (IC) e no Instituto do Câncer do Estado de São Paulo (ICESP), do Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (HCFMUSP) que foram submetidos a USER-3D pré-operatório. Os parâmetros analisados por meio do USER-3D foram comparados aos achados da anatomia patológica do espécime cirúrgico obtido após o procedimento. Os exames de USER-3D foram realizados pelo mesmo médico, sendo cego dos outros métodos diagnósticos e dos achados da patologia, a qual também não tinha os resultados do estadiamento clínico pré-operatório. Todos os pacientes foram submetidos a enema evacuatório no dia anterior e na manhã do exame. Foram avaliados a sensibilidade, especificidade, valores preditivos positivo e negativo, área sobre a curva e o índice Kappa do USER-3D em comparação ao anatomopatológico, considerado exame padrão ouro. O coeficiente de correlação intraclasse (CCI) foi utilizado para analisar a extensão e porcentagem de acometimento da lesão na parede retal. RESULTADOS: No período de 3 anos, 44 pacientes foram estudados, 27 mulheres, com idade média de 63,5 anos. Houve 12 lesões benignas e 32 malignas de reto, sendo 30 submetidas à ressecção local e 14 à radical. O USER-3D determinou a diferenciação da profundidade de invasão tumoral na submucosa com sensibilidade de 77,3% (CI95% - 54,6%-92,2%), especificidade de 86,4% (CI95% - 65.1%-97.1%), valor preditivo positivo de 85% (CI95% - 62,1%-96,8%), valor preditivo negativo de 79,2% (CI95% - 57,8%-92,9%) e área sobre a curva de 0,82% (CI95% - 0,7%-0,96%). O índice Kappa ponderado para profundidade de invasão da parede retal (T) foi de 0,672 (IC95%: 0,493; 0,850), considerado grau de concordância substancial. Para o envolvimento linfonodal (N) não houve concordância adequada entre o USER-3D e o anatomopatológico (k=-0,164) nos 14 casos analisados. A CCI calculada para extensão da lesão foi moderada (0,45) para a extensão em centímetros, mas adequada (0,66) para porcentagem de envolvimento da circunferência. O gráfico de Bland-Altman mostrou que lesões com extensão de até 5 cm e 50% de acometimento têm melhor correlação com o espécime cirúrgico. CONCLUSÕES: USER-3D foi eficaz para a determinação da invasão da parede retal, sendo seguro na determinação da extensão de lesões até 5 cm e porcentagem de acometimento da circunferência até 50%. O método mostrou baixa eficácia na avaliação linfonodal, em subgrupo limitado de pacientes / INTRODUCTION: Loco-regional staging and treatment of extra peritoneal rectal neoplasms is still a controversial subject. There is no perfect method, substantially sensitive and specific for staging rectal wall invasion and lymph node involvement. Adequate oncologic staging of rectal neoplasias has major importance in both, treatment and prognostic evaluation. Therefore, the use of supplementary diagnostic methods such as endorectal ultrasound (ERUS) and magnetic resonance imaging (MRI) of the pelvis can promote an accurate assessment of tumor invasion in the rectal wall and lymph node involvement. OBJECTIVE: To correlate the findings of three-dimensional (3D) ERUS with pathology specimen of extra peritoneal rectal neoplasia referred directly to surgery after diagnosis, in regards to depth of rectal wall invasion (T), lymph node involvement, percentage of circumferential rectal wall invasion and tumor extension. METHODS: A prospective study was performed in patients with middle and distal rectal tumors followed at University of São Paulo, School of Medicine and Cancer Institute of State of São Paulo (ICESP), who underwent 3D-ERUS for preoperative evaluation. The parameters analyzed with 3D-ERUS were compared with pathology findings of the surgical specimen obtained after the procedure. A single doctor who performed the exams reported the 3D-ERUS studies and was blind of other methods as well as the pathologic findings. All patients underwent retrograde bowel enema the day before and in the morning of the test. The authors evaluated sensitivity, specificity, positive and negative predictive values, area under curve and kappa index of the 3D-ERUS as compared to pathologic findings, considered the gold standard. For extension and percentage of tectal wall involvement intraclass correlation index was applied. RESULTS: At 3-years period, 44 patients were studied, 27 females, with a mean age of 63.5 years, who had 12 rectal adenomas and 32 adenocarcinomas and underwent local resection (30) or radical resection (14). Value for 3D-ERUS to determine depth of rectal wall invasion sensitivity was 77.3% (CI95% - 54.6%-92.2%), specificity was 86.4% (CI95% - 65.1%- 97.1%), positive predictive value was 85% (CI95% - 62.1%- 96.8%), negative predictive value was 79.2% (CI95% - 57.8%-92.9%) and area under curve was 0,82% (CI95% - 0.7%-0.96%). The weighted kappa index for the depth of invasion in the rectal wall (T) evaluation was 0.67 (IC95%: 0.49; 0.85), considered substantial agreement. For N involvement there was any agreement between 3D-ERUS and histopathology, with K=-0.164. Intraclass correlation was calculated for lesion extension and was moderate (0.45) for extension in centimeters and adequate (0.66) for percentage of circumference involvement. A Bland-Altman graph was performed and showed that tumor extensions until 5 cm and 50% of wall involvement have a good correlation to specimen size. CONCLUSION: 3D-ERUS was effective for determining rectal wall invasion and evaluation of extension of lesions until 5 cm and 50% of rectal wall involvement. However, this method showed a lack of efficacy for evaluation of lymph node involvement for early rectal tumors in this limited subset of patients

INTRODUÇÃO: O câncer do reto médio e distal ainda é assunto bastante controverso, especialmente no que se refere ao estadiamento locorregional e opções terapêuticas. Busca-se um método sensível e específico para a avaliação da profundidade de invasão da parede retal e o envolvimento linfonodal. O adequado estadiamento da neoplasia do reto extraperitoneal é de suma importância no manejo terapêutico e prognóstico do paciente. Diversos métodos têm sido descritos para a avaliação da disseminação locorregional das neoplasias do reto, que variam desde o toque retal até a ressonância magnética da pelve e a ultrassonografia endorretal bi e tridimensional. OBJETIVO: Correlacionar os achados da profundidade de invasão tumoral na parede retal (T), comprometimento linfonodal (N), extensão e porcentagem de acometimento da lesão à ultrassonografia endorretal tridimensional (USER-3D) com o exame anatomopatológico (AP) de pacientes portadores de neoplasia de reto extraperitoneal submetidos a procedimento cirúrgico após o diagnóstico e estadiamento clínico prévio. MÉTODO: Estudo prospectivo foi realizado com pacientes portadores de neoplasia de reto médio e distal seguidos no Instituto Central (IC) e no Instituto do Câncer do Estado de São Paulo (ICESP), do Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (HCFMUSP) que foram submetidos a USER-3D pré-operatório. Os parâmetros analisados por meio do USER-3D foram comparados aos achados da anatomia patológica do espécime cirúrgico obtido após o procedimento. Os exames de USER-3D foram realizados pelo mesmo médico, sendo cego dos outros métodos diagnósticos e dos achados da patologia, a qual também não tinha os resultados do estadiamento clínico pré-operatório. Todos os pacientes foram submetidos a enema evacuatório no dia anterior e na manhã do exame. Foram avaliados a sensibilidade, especificidade, valores preditivos positivo e negativo, área sobre a curva e o índice Kappa do USER-3D em comparação ao anatomopatológico, considerado exame padrão ouro. O coeficiente de correlação intraclasse (CCI) foi utilizado para analisar a extensão e porcentagem de acometimento da lesão na parede retal. RESULTADOS: No período de 3 anos, 44 pacientes foram estudados, 27 mulheres, com idade média de 63,5 anos. Houve 12 lesões benignas e 32 malignas de reto, sendo 30 submetidas à ressecção local e 14 à radical. O USER-3D determinou a diferenciação da profundidade de invasão tumoral na submucosa com sensibilidade de 77,3% (CI95% - 54,6%-92,2%), especificidade de 86,4% (CI95% - 65.1%-97.1%), valor preditivo positivo de 85% (CI95% - 62,1%-96,8%), valor preditivo negativo de 79,2% (CI95% - 57,8%-92,9%) e área sobre a curva de 0,82% (CI95% - 0,7%-0,96%). O índice Kappa ponderado para profundidade de invasão da parede retal (T) foi de 0,672 (IC95%: 0,493; 0,850), considerado grau de concordância substancial. Para o envolvimento linfonodal (N) não houve concordância adequada entre o USER-3D e o anatomopatológico (k=-0,164) nos 14 casos analisados. A CCI calculada para extensão da lesão foi moderada (0,45) para a extensão em centímetros, mas adequada (0,66) para porcentagem de envolvimento da circunferência. O gráfico de Bland-Altman mostrou que lesões com extensão de até 5 cm e 50% de acometimento têm melhor correlação com o espécime cirúrgico. CONCLUSÕES: USER-3D foi eficaz para a determinação da invasão da parede retal, sendo seguro na determinação da extensão de lesões até 5 cm e porcentagem de acometimento da circunferência até 50%. O método mostrou baixa eficácia na avaliação linfonodal, em subgrupo limitado de pacientes / INTRODUCTION: Loco-regional staging and treatment of extra peritoneal rectal neoplasms is still a controversial subject. There is no perfect method, substantially sensitive and specific for staging rectal wall invasion and lymph node involvement. Adequate oncologic staging of rectal neoplasias has major importance in both, treatment and prognostic evaluation. Therefore, the use of supplementary diagnostic methods such as endorectal ultrasound (ERUS) and magnetic resonance imaging (MRI) of the pelvis can promote an accurate assessment of tumor invasion in the rectal wall and lymph node involvement. OBJECTIVE: To correlate the findings of three-dimensional (3D) ERUS with pathology specimen of extra peritoneal rectal neoplasia referred directly to surgery after diagnosis, in regards to depth of rectal wall invasion (T), lymph node involvement, percentage of circumferential rectal wall invasion and tumor extension. METHODS: A prospective study was performed in patients with middle and distal rectal tumors followed at University of São Paulo, School of Medicine and Cancer Institute of State of São Paulo (ICESP), who underwent 3D-ERUS for preoperative evaluation. The parameters analyzed with 3D-ERUS were compared with pathology findings of the surgical specimen obtained after the procedure. A single doctor who performed the exams reported the 3D-ERUS studies and was blind of other methods as well as the pathologic findings. All patients underwent retrograde bowel enema the day before and in the morning of the test. The authors evaluated sensitivity, specificity, positive and negative predictive values, area under curve and kappa index of the 3D-ERUS as compared to pathologic findings, considered the gold standard. For extension and percentage of tectal wall involvement intraclass correlation index was applied. RESULTS: At 3-years period, 44 patients were studied, 27 females, with a mean age of 63.5 years, who had 12 rectal adenomas and 32 adenocarcinomas and underwent local resection (30) or radical resection (14). Value for 3D-ERUS to determine depth of rectal wall invasion sensitivity was 77.3% (CI95% - 54.6%-92.2%), specificity was 86.4% (CI95% - 65.1%- 97.1%), positive predictive value was 85% (CI95% - 62.1%- 96.8%), negative predictive value was 79.2% (CI95% - 57.8%-92.9%) and area under curve was 0,82% (CI95% - 0.7%-0.96%). The weighted kappa index for the depth of invasion in the rectal wall (T) evaluation was 0.67 (IC95%: 0.49; 0.85), considered substantial agreement. For N involvement there was any agreement between 3D-ERUS and histopathology, with K=-0.164. Intraclass correlation was calculated for lesion extension and was moderate (0.45) for extension in centimeters and adequate (0.66) for percentage of circumference involvement. A Bland-Altman graph was performed and showed that tumor extensions until 5 cm and 50% of wall involvement have a good correlation to specimen size. CONCLUSION: 3D-ERUS was effective for determining rectal wall invasion and evaluation of extension of lesions until 5 cm and 50% of rectal wall involvement. However, this method showed a lack of efficacy for evaluation of lymph node involvement for early rectal tumors in this limited subset of patients

Contexte: Le cancer colo-rectal est la deuxième cause de décès, par ordre de fréquence. L’utilisation de l’imagerie dans la stadification du cancer colo-rectal est un élément important de la prise en charge de la maladie. L’échographie endoscopique est une modalité qui permet de préciser la profondeur de l’atteinte néoplasique. Les données probantes concernant la performance diagnostique dans l’identification de cancers peu avancés sont variables. Objectif : Effectuer une revue systématique sur la performance diagnostique de l’échographie endoscopique dans l’identification de cancer de stade T1 et T2. Devis : Revue systématique. Sources bibliographiques : PubMed, EMBASE, Ovid and Cochrane library Méthodes: Dans un premier temps, une recherche de revue systématique publiée dans les 15 dernières années fût effectuée sur la précision diagnostique de l’échographie endoscopique dans les banques PubMed, Cochrne et trip database. Deux revues systématiques, publiées en 2008 et 2009 fûrent identifiées. Une deuxième recherche portant sur des études primaires a été effectuée pour la période 2009 à 2016, dans les mêmes banques bibliographiques. La qualité des études primaires a été évaluée à l’aide de la grille QUADAS2. Les mots clés utilisés étaient échographie endoscopique, EUS, cancer rectal, histo-pathologie, staging. Sélection d’études : Les critères d’inclusion : population adulte avec diagnostic de cancer du rectum pas avancé, articles complets publiés dans des revues avec comité de pairs, articles en anglais. Critères d’exclusion : population pédiatrique, cancers avancés avec atteinte métastatique, patients évalués avec d’autres modalités (CT ou IRM) sans échographie endoscopique, absence de confirmation histologique. Résultats : Dix articles, publiés depuis 2009, répondaient aux critères d’inclusion. Ces articles furent ajoutés aux articles retenus dans les revues systématiques déjà publiées. Au total,49 articles sont inclus dans cette revue systématique. La performance diagnostique de l’échographie endoscopique a été évaluée en calculant la sensitivité et la spécificité des études regroupées. Pour le stade T1, les valeurs de sensitivité et spécificité étaient 0.84 (CI 0.75-0.91) et 0.93 (CI 0.86–0.97), respectivement. Pour le stade T2 les valeurs de sensitivité et spécificité étaient 0.83 (CI 0.74–0.90) et 0.93 (CI 0.86–0.97), respectivement. Conclusion: L’échographie endoscopique présente une performance diagnostique pour l’identification de cancers de stade T1 et T2. Ceci permet d’orienter des patients vers des chirurgies moins invasives avec une survie égale et un taux de complications inférieures comparativement à des chirurgies plus invasives. / Background: Colorectal cancer (CRC) is the second leading cause of death. The use of preoperative imaging in the staging of (CRC) plays a major role in the management. Endorectal ultrasound (ERUS) is a precise imaging modality to determine the depth of penetration. The data on the precision of (ERUS) to predict early stage of rectal cancer has been variable Objectives: To conduct a systematic review, on the diagnostic performance of (ERUS) in the staging of T1 and T2 CRC. Design: Systematic review. Data sources: A literature search via PubMed, EMBASE, Ovid and Cochrane library. METHODS: An initial search for systematic review articles published in the last 15 years on the diagnostic accuracy of EUS in the staging of CRC using PubMed, Cochrane library, and trip database was conducted. After finding two systematic reviews that were published in 2008 and 2009, a second search of original studies published since the systematic reviews were conducted using the same databases from 2009 to 2016. The primary studies included in the systematic reviews and the primary studies published afterwards were included in the review. Methodological quality was applied using a modified version of the quality assessment of diagnostic accuracy studies (QUADAS2) tool. Terms used for search were endoscopic ultrasound, EUS, rectal cancer, histo-pathological finding, and staging. Study selection: Inclusion criteria includes adult people diagnosed with early stage CRC, all articles in english language and must be a full manuscripts published in peer-reviews journals. Exclusion criteria includes any recurrent or metastasis cancer and children with rectal cancer. Patients who were staged preoperatively by other imaging modality (MRI or CT) and no comparison with post operative pathology. Results: The search identified 420 articles, 97 articles were duplicate and excluded, and 232 refined articles were screened for title and abstract, reviewed. Thirty-two full text studies were assessed for eligibility, and ten published as full text and met the inclusion criteria; they were added to the articles identified in the earlier systematic reviews a total of 49 articles. Results of the evaluation of the accuracy of ERUS analyzed according to the diagnostic measures of sensitivities and specificities calculated for each study. The pooled sensitivity and specificity of EUS for stage T1 CRC was 0.84 (CI 0.75-0.91) and 0.93 (CI 0.86–0.97), and for T2 was 0.83 (CI 0.74–0.90) and 0.93(CI 0.86–0.97) respectively. Conclusion: The range of sensitivity and specificity values suggest that EUS performs well in accurately staging T1 and T2 cancers. Further advancement in this technology will lead to an improved diagnosis, clinical decision-making, and reduce the over staging drawback.