The Reproducibility Index of Pathological Diagnosis and Rare Cases. The Results of the On-line Diagnostic Competition “Final Diagnosis”

Abstract

Introduction/ Background

UNIM Ltd. have created the SAAS platform DPathology that can be used for saving and studying histological slides and it doesn’t require an installation of a special software. You can use the platform with all the modern internet browsers. The SAAS platform gives all the specialists a chance to analyze remotely digital histological slides. It increases the accuracy of diagnostics and speeds up the medical assessment

Aims

To indicate the importance of collecting rare cases and expert assessment via digital microscopy

Using the Digital Pathology© platform to carry out educational and competitive diagnostic measures.

Methods

Fourteen rare cases from different sub-specializations field in pathology were selected by UNIM LTD with expert’s pathologists from the Czech Republic and Italy and additionally validated in Norwayand theUK (blind method). The slides were digitized and introduced withclinical information to 250 specialists registered to take part in the competition “Final diagnosisâ€Â©.

Results

The range of the totally correct answers varies between 3 and 56 percent. The most difficult case for the participants was the one with no tumorous pathology: ectopic hamartomatous thymoma [1]. There were 3 percent of full match. The biggest number of full match to experts’ diagnoses can be seen in the case: Grade 2 central chondrosarcoma with 72% of agreement. To analyze the disagreements we divided them in two groups:

Mayor disagreement – potentially not correct histological diagnosis will change the clinical tactics of patient’s treatment (considering the malignant pathology as a benign pathology, considering the benign pathology as a malignant, changing the stage of disease). Potentially wrong pathological diagnosis leads to wrong course

of patient’s treatment and wrong chemotherapy, etc. Minor disagreement – potentially incorrect diagnosis doesn’t have any clinical matter. This tactic showed that the case of hyalinized endometrioid adenocarcinoma [ [2] turned out to be the most difficult one for the participants. The range of mayor disagreement here was 66 percent, mostly because cases was interpret as carcinosarcoma (63/93), while the agreement is 14 percent. And myxoinflammatory fibroblastic sarcoma of soft tissues [ 3], with major disagreement in 67.4% (pict case 10).

All the data is shown in table 1.

Â	Â	Agreement %	Minor disagreement %	Major disagreement %
case1, N=97	Teratocarcinosarcoma of the nasal cavity [4]	0,12	0,61	0,26
case 2, N=89	Juxtaoral organ of Chievitz [5]	0,13	0,50	0,37
case3, N=93	Mammary Analogue Secretory Carcinoma of Salivary Glands, Containing the ETV6-NTRK3 Fusion Gene [6]	0,365	0,48	0,15
case 4, N=96	Low-grade sebaceous carcinoma of the skin [7]	0,10	0,68	0,23
case 5, N=99	t(6;11) translocation carcinoma (Ro- sette-forming tumor of the kidney) [8]	0,31	0,35	0,32
case 6, N=93	Hyalinized endometrioid adenocarcinoma [9]	0,14	0,20	0,67
case 7, N=96	Ectopic hamartomatous thymoma	0,03	0,82	0,15
case 8, N=91	Prolapse of the fallopian tube after hyster- ectomy	0,16	0,67	0,14
case 9, N=87	Phosphaturic mesenchymal tumor of soft tissues with calcification	0,23	0,34	0,33
case 10, N=92	Myxoinflammatory fibroblastic sarcoma of soft tissues [3]	0,13	0,20	0,67
case 11, N=87	Enchondroma	0,35	0,15	0,49
case12, N=88	Atypical chondromatous tumor/ Grade 1 chondrosarcoma	0,20	0,42	0,375
case 13, N=87	Grade 2 central chondrosarcoma	0,72	0,07	0,21