Locoregional staging using magnetic resonance imaging in rectal cancer: influence of diffusion-weighted imaging in the diagnostic accuracy.

OBJECTIVE: To assess whether the accuracy in locoregional staging using magnetic resonance imaging (MRI) in rectal cancer (primary or post-chemoradiotherapy) improves by adding diffusion-weighted imaging, according to the radiologist's degree of experience. METHOD: Retrospective study on 100 MRI records (1.5 T, 2011-2016) from patients with rectal cancer (reference standard: histology of surgical specimens). Ten radiologists (three experienced in rectal cancer, three specialized in other areas and four residents) individually reviewed each case twice: first, evaluating just high-resolution T2-weighted sequences; second, evaluation of diffusion-weighted plus high-resolution ones. The analysis focused on the differentiation between early (0-I) and advanced (II-IV) stages. Accuracy, sensitivity/specificity and predictive values were calculated. RESULTS: Experienced radiologists showed some worsening by adding diffusion-weighted imaging, mainly at primary staging (accuracy: 0.769 to 0.701). Inexperienced radiologists presented a post-chemoradiotherapy improvement (accuracy: 0.574 to 0.642; specificity of 19.1 to 29.8%), although with no other remarkable changes. Residents demonstrated a worsening at primary staging by adding diffusion (accuracy: 0.670 to 0.633; specificity: 45.8 to 39.6%), but post-chemoradiotherapy improvement (sensitivity: 80.6 to 87%). The differences between both reviews were not statistically significant. CONCLUSIONS: No significant differences were found in the distinction between early and advanced rectal tumors secondary to adding diffusion-weighted imaging to high-resolution T2-weighted sequences.

OBJETIVO: Evaluar si la eficacia en la estadificación locorregional por resonancia magnética (RM) del cáncer de recto (primaria o posneoadyuvancia) mejora al añadir imágenes potenciadas en difusión, según la experiencia previa del radiólogo. MÉTODO: Estudio retrospectivo sobre 100 RM de 1.5 T (2011-2016) de pacientes con cáncer rectal (estándar de referencia: estadiaje histológico de pieza quirúrgica). Diez radiólogos (tres con experiencia en cáncer rectal, tres inexpertos y cuatro residentes) evaluaron individualmente cada caso dos veces: primero, solo secuencias T2 de alta resolución; segundo, valoración conjunta con difusión. Se analizó la diferenciación entre estadios precoces (0-I) y avanzados (II-IV), y se calcularon la precisión, la sensibilidad y la especificidad, y los valores predictivos. RESULTADOS: Al agregar la difusión, los radiólogos experimentados presentaron peores resultados, sobre todo en estadiaje primario (precisión: 0.769 a 0.701). Los inexpertos mostraron mejoría posneoadyuvancia (precisión: 0.574 a 0.642; especificidad: 19.1 a 29.8%), sin otros cambios destacables. Los residentes manifestaron peores resultados en estadiaje primario (precisión: 0.670 a 0.633; especificidad: 45.8-39.6%), pero mejoría posneoadyuvancia (sensibilidad: 80.6 a 87%). Las diferencias entre ambas revisiones no fueron estadísticamente significativas. CONCLUSIONES: No se encontraron diferencias significativas en la distinción entre tumores rectales precoces y avanzados al añadir secuencias de difusión al uso de secuencias T2 de alta resolución.

Primary and post-chemoradiotherapy MRI detection of extramural venous invasion in rectal cancer: the role of diffusion-weighted imaging.

OBJECTIVE: To assess the added value of diffusion-weighted imaging (DWI) to high-resolution T2-weighted sequences (HRT2w) in MRI detection of extramural venous infiltration (EMVI), according to the previous experience of radiologists. METHODS: A cross-sectional study of a retrospective database including 1.5 T MRI records from 100 patients with biopsy-proven rectal cancer (2011-2016; 75 male/25 female, average 63 y/o), which included primary staging (54) and post-chemoradiotherapy follow-up MRIs (46). The reference standard was histology of surgical specimens. All cases were individually blindly reviewed by ten radiologists: three specialists in abdominal radiology, three specialized in different areas and four residents. In each case, the presence of EMVI was assessed twice: first, using just HRT2w; second, with DWI added to HRT2w. The results were pooled by experience, analyzing sensitivity, specificity, accuracy (area under ROC curve), likelihood ratios, predictive values and overstaging/understaging. RESULTS: Addition of DWI improved diagnostic performance by specialists radiologists, particularly post-chemoradiotherapy (accuracy 0.74-0.84; positive likelihood ratio 3.9-9.1; overstaging 16-8%), less so at primary staging (specificity 76-87.2%; overstaging 21-11%). Non-specialist radiologists also improved, but only at primary staging (accuracy 0.59-0.63). Residents showed small changes, except for notably increased sensitivity in both primary staging (35.7-43%) and post-chemoradiotherapy (41.7-58.3%) staging, at the expense of increased overstaging. CONCLUSIONS: The addition of DWI improved the diagnostic performance of EMVI by experienced radiologists, downgrading overstaging, especially in post-chemoradiotherapy follow-up. It resulted in fewer changes for inexperienced radiologists (enhanced primary staging) and residents (increased sensitivity).

Primary and post-chemoradiotherapy staging using MRI in rectal cancer: the role of diffusion imaging in the assessment of perirectal infiltration.

PURPOSE: To analyze changes in MRI diagnostic accuracy in main rectal tumor (T) evaluation resulting from the use of diffusion-weighted imaging (DWI), according to the degree of experience of the radiologist. METHODS: This is a cross-sectional study of a database including one hundred 1.5 T MRI records (2011-2016) from patients with biopsy-proven rectal cancer, including primary staging and post-chemoradiotherapy follow-up. All cases were individually blindedly reviewed by ten radiologists: three experienced in rectal cancer, three specialized in other areas, and four residents. Each case was assessed twice to detect perirectal infiltration: first, evaluating just high-resolution T2-weighted sequences (HRT2w); second, evaluation of DWI plus HRT2w sequences. Results were pooled by experience, calculating accuracy (area under ROC curve), sensitivity and specificity, predictive values, likelihood ratios, and overstaging/understaging. Histology of surgical specimens provided the reference standard. RESULTS: DWI significantly improved specificity by experienced radiologists in primary staging (63.2% to 75.9%) and, to a lesser extent, positive likelihood ratio (2.06 to 2.87); minimal changes were observed post-chemoradiotherapy, with a slight decrease of accuracy (0.657 to 0.626). Inexperienced radiologists showed a similar pattern, but with slight enhancement post-chemoradiotherapy (accuracy 0.604 to 0.621). Residents experienced small changes, with increased sensitivity/decreased specificity in both primary (69% to 72%/67.2% to 64.7%) and post-chemoradiotherapy (68.1% to 73.6%/47.3% to 44.6%) staging. CONCLUSIONS: Adding DWI to HRT2w significantly improved specificity for the detection of perirectal infiltration at primary staging by experienced radiologists and also by inexperienced ones, although to a lesser extent. In the post-neoadjuvant treatment subgroup, only minimal changes were observed.

Computer vision techniques for breast tumor ultrasound analysis.

In this paper, we present a new approach to the segmentation and analysis of solid breast nodules in ultrasonography. We have applied computer vision techniques to segment the nodules and analyze a series of diagnostic criteria which can help discriminate malignant and benignant tumors. The segmentation is carried out in a semiautomatic way, whereas the analysis of the diagnostic criteria involves several computational methods. The techniques which we propose have provided quite satisfactory results and show the usefulness of image processing in the diagnosis through medical imaging.