Classificação de Bosniak de cistos renais versão 2005 e 2019: um estudo de reprodutibilidade
Data
2022-04
Autores
Pacheco, Eduardo Oliveira 
Orientadores
D'Ippolito, Giuseppe
Tipo
Dissertação de mestrado
Título da Revista
ISSN da Revista
Título de Volume
Resumo
Objetivo: Em uma condição tão prevalente como os cistos renais, as classificações devem apresentar reprodutibilidade satisfatória não apenas entre leitores subespecializados. Nosso objetivo foi comparar a concordância interobservador da classificação de Bosniak versão 2019 (CB19) e a classificação de Bosniak versão 2005 (CB05) para leitores não subespecializados na TC e RM, e verificar se a CB19 leva a um rebaixamento dos cistos renais quando comparada à CB05 e a um aumento de confiança na sua categorização. Métodos: Avaliamos 50 cistos renais em 47 pacientes na RM e TC (25 por método). Dezoito leitores (nove residentes de radiologia do terceiro ano e nove fellows de imagem abdominal) avaliaram as imagens usando a CB19 e a CB05 com um intervalo de oito semanas. A estatística Kappa foi usada para avaliar a concordância interobservador. Uma pontuação média das categorias Bosniak em todos os leitores avaliou se houve rebaixamento das lesões na CB19. Resultados: Os maiores valores de Kappa foram encontrados para os fellows na CB05-RM (κ = 0,51), e os menores valores foram encontrados para os residentes na CB05-RM e fellows na CB19-RM (ambos κ = 0,36). Na CB05, a concordância interobservador foi moderada para RM e TC (κ = 0,42 e 0,43, respectivamente), enquanto na CB19, foi razoável (κ = 0,38 e 0,40, respectivamente). As melhores concordâncias ocorreram nas categorias I (κ = 0,49− 0,69) e IV (κ = 0,45− 0,51). As concordâncias mais fracas ocorreram na categoria IIF (κ = 0,18 na CB19-CT). Houve um aumento mediano moderado de CB05 para CB19 em termos de categorias Bosniak para ambos os métodos (RM [Z=-2,058, p = 0,040] e TC [Z=-2,509, p = 0,012]). Conclusões: a CB19, quando comparada com a CB05, não melhorou a concordância interobservador, não diminuiu a proporção de cistos categorizados em classes Bosniak mais altas e não aumentou o grau de confiança entre os leitores.
Objective: In a condition so prevalent as renal cysts, classifications should display satisfactory reproducibility not only among subspecialized readers. We aimed to compare the interobserver agreement of the Bosniak classification version 2019 (CB19) and current Bosniak classification (CB05) for non-subspecialized readers on CT and MRI and to verify whether CB19 leads to a downgrade of renal cystic masses when compared to CB05. Methods: We evaluated 50 renal cystic masses in 47 patients on MR and CT (25 per method). Eighteen readers (nine third-year radiology residents and nine abdominal imaging fellows) assessed the images using CB19 and CB05 with an eight-week interval. Kappa statistic was used to assess agreement. An average score of Bosniak categories across all raters evaluated if there was downgrading of lesions on CB19. Results: The highest values of Kappa were found for fellows on CB05-MR (κ = 0.51), and the lowest values were found for residents on CB05-MR and fellows on CB19-MR (both κ = 0.36). On CB05, interobserver agreement was moderate for MR and CT (κ = 0.42 and 0.43, respectively), whereas on CB19, it was fair (κ = 0.38 and 0.40, respectively). The best agreements were in categories I (κ = 0.49− 0.69) and IV (κ = 0.45− 0.51). The poorest agreements occurred at IIF (κ = 0.18 on CB19-CT). There was a moderate median increase from CB05 to CB19 in terms of Bosniak categories for both methods (MR [Z=-2.058, p = 0.040] and CT [Z=- 2.509, p = 0.012]). Conclusions: CB19, when compared to CB05, did not improve interobserver agreement nor diminished the proportion of masses categorized into higher Bosniak classes among non-subspecialized readers.
Objective: In a condition so prevalent as renal cysts, classifications should display satisfactory reproducibility not only among subspecialized readers. We aimed to compare the interobserver agreement of the Bosniak classification version 2019 (CB19) and current Bosniak classification (CB05) for non-subspecialized readers on CT and MRI and to verify whether CB19 leads to a downgrade of renal cystic masses when compared to CB05. Methods: We evaluated 50 renal cystic masses in 47 patients on MR and CT (25 per method). Eighteen readers (nine third-year radiology residents and nine abdominal imaging fellows) assessed the images using CB19 and CB05 with an eight-week interval. Kappa statistic was used to assess agreement. An average score of Bosniak categories across all raters evaluated if there was downgrading of lesions on CB19. Results: The highest values of Kappa were found for fellows on CB05-MR (κ = 0.51), and the lowest values were found for residents on CB05-MR and fellows on CB19-MR (both κ = 0.36). On CB05, interobserver agreement was moderate for MR and CT (κ = 0.42 and 0.43, respectively), whereas on CB19, it was fair (κ = 0.38 and 0.40, respectively). The best agreements were in categories I (κ = 0.49− 0.69) and IV (κ = 0.45− 0.51). The poorest agreements occurred at IIF (κ = 0.18 on CB19-CT). There was a moderate median increase from CB05 to CB19 in terms of Bosniak categories for both methods (MR [Z=-2.058, p = 0.040] and CT [Z=- 2.509, p = 0.012]). Conclusions: CB19, when compared to CB05, did not improve interobserver agreement nor diminished the proportion of masses categorized into higher Bosniak classes among non-subspecialized readers.