Machine Learning-Based Integration of Prognostic Magnetic Resonance Imaging Biomarkers for Myometrial Invasion Stratification in Endometrial Cancer

BACKGROUND

Estimation of the depth of myometrial invasion (MI) in endometrial cancer is pivotal in the preoperatively staging. Magnetic resonance (MR) reports suffer from human subjectivity. Multiparametric MR imaging radiomics and parameters may improve the diagnostic accuracy.

PURPOSE

To discriminate between patients with MI ≥ 50% using a machine learning-based model combining texture features and descriptors from preoperatively MR images.

STUDY TYPE

Retrospective.

POPULATION

One hundred forty-three women with endometrial cancer were included. The series was split into training (n = 107, 46 with MI ≥ 50%) and test (n = 36, 16 with MI ≥ 50%) cohorts.

FIELD STRENGTH/SEQUENCES

Fast spin echo T2-weighted (T2W), diffusion-weighted (DW), and T1-weighted gradient echo dynamic contrast-enhanced (DCE) sequences were obtained at 1.5 or 3 T magnets.

ASSESSMENT

Tumors were manually segmented slice-by-slice. Texture metrics were calculated from T2W and ADC map images. Also, the apparent diffusion coefficient (ADC), wash-in slope, wash-out slope, initial area under the curve at 60 sec and at 90 sec, initial slope, time to peak and peak amplitude maps from DCE sequences were obtained as parameters. MR diagnostic models using single-sequence features and a combination of features and parameters from the three sequences were built to estimate MI using Adaboost methods. The pathological depth of MI was used as gold standard.

STATISTICAL TEST

Area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, accuracy, positive predictive value, negative predictive value, precision and recall were computed to assess the Adaboost models performance.

RESULTS

The diagnostic model based on the features and parameters combination showed the best performance to depict patient with MI ≥ 50% in the test cohort (accuracy = 86.1% and AUROC = 87.1%). The rest of diagnostic models showed a worse accuracy (accuracy = 41.67%-63.89% and AUROC = 41.43%-63.13%).

DATA CONCLUSION

The model combining the texture features from T2W and ADC map images with the semi-quantitative parameters from DW and DCE series allow the preoperative estimation of myometrial invasion. EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: Stage 3.

Ultrastaging protocol in sentinel lymph node for apparent early stage ovarian cancer

OBJECTIVE

The objective of the present study is to determine the role of sentinel lymph node (SLN) ultrastaging in apparent early-stage ovarian cancer.

METHODS

We previously demonstrated the feasibility of SLN in early-stage ovarian cancer in a pilot study and in a clinical trial (NCT03452982). The SLN of the 30 patients involved in both were processed following an ultrastaging protocol. The cost of ultrastaging processing was also reported.

RESULTS

A SLN was detected in up to 91.3% and 90% in the pelvic and para-aortic region, respectively. In all cases, a SLN was detected at least in one field, pelvic or para-aortic. The mean time from injection to SLN resection was 53.3 ± 20.3 min. Two of 30 (6.6%) patients had a contralateral SLN in the para-aortic field, but no patients had contralateral SLN within the pelvic field after injection. The mean number of harvested SLN was 2.1 ± 1.4 (range: 0-5) and 2.7 ± 1.5 (range: 0-7) in the pelvic and para-aortic region, respectively. Two patients were upgraded to stage IIIA1 because of lymph node metastasis. In the first case, based on single sections and haematoxylin and eosin (H&E) examination, a pelvic SLN micrometastasis (1 mm) was found on the first H&E section. By using the ultrastaging protocol, the size of the metastasis was increased to 2.1 mm (macrometastasis). In the same patient, the ultrastaging study of the inframesenteric para-cava SLNs found isolated tumour cells in the subcapsular and interfollicular lymph nodes sinus in one of the two SLN harvested (in one of the sections at the fourth and fifth ultrastage levels). The other upstaged case was a para-aortic macrometastasis in a patient in whom the SLN was not identified in the para-aortic field because of the absence of migration from the infundibulo-pelvic stump injection. The cost of ultrastaging in each patient depended on the total number of SLN retrieved, averaging 96.8 € (range: 0-230.5) and 124.5 € (range: 0-322.7€) for pelvic and para-aortic SLN, respectively.

CONCLUSIONS

A uniform protocol for ultrastaging is essential for lower-volume metastasis detection and to provide reproducible information between upcoming studies, as evidence about SLN in ovarian cancer is growing.

Vincristine, an efficacious alternative for diffuse neonatal haemangiomatosis

Diffuse neonatal haemangiomatosis (DNH) is an uncommon condition characterized by multiple cutaneous and visceral haemangiomas frequently causing severe complications. Corticosteroids constitute the first therapeutic line; however, when they fail, other alternatives are available, provided possible side effects are closely monitored during and after treatment. We present a case of life-threatening DNH, non-responsive to corticosteroids, successfully treated with Vincristine with minor side effects. We conclude that Vincristine is a valid alternative in corticosteroid-resistant DNH.