06 June 2023 : Database Analysis
Enhancing Renal Tumor Detection: Leveraging Artificial Neural Networks in Computed Tomography Analysis
Mateusz Glembin
1ABCDEFG*, Aleksander Obuchowski 2ABCDEG, Barbara Klaudel 3ABCDEG, Bartosz Rydziński 2ABCDEG, Roman Karski 2ABCDEG, Paweł Syty 24ABCDEFG, Patryk Jasik 24ABCDEFG, Wojciech Józef Narożański1AG
DOI: 10.12659/MSM.939462
Med Sci Monit 2023; 29:e939462
Table 1 Characteristics of collected tumors and dataset statistics.
| Tumor type | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ccRCC | pRCC | chRCC | Other malignant | Total malignant | AML | Oncocytoma | Other benign | Total benign | Total | ||
| All collected tumors | Number of tumors (%) | 201 (56.3) | 38 (10.6) | 25 (7.0) | 1 (0.3) | 265 (74.2) | 68 (19.0) | 20 (5.6) | 4 (1.1) | 92 (25.8) | 357 (100) |
| Average size ±SD (mm) | 46.8± 27.6 | 38.4± 22.8 | 63.8± 36.3 | 61± 0 | 47.3± 28.4 | 11.7± 9.2 | 29.3± 16.3 | 18.0± 8.8 | 15.8± 13.2 | 39.2± 28.9 | |
| Number of collected arterial-phase images (DICOM images, %) | 4757 (66) | 717 (9.9) | 999 (13.9) | 16 (0.2) | 6489 (90.0) | 280 (3.9) | 411 (5.7) | 27 (0.4) | 718 (10.0) | 7207 (100) | |
| Test dataset | Number of tumors (%) | 22 (57.9) | 4 (10.5) | 2 (5.3) | – | 28 (73.7) | 8 (21.1) | 2 (5.3) | – | 10 (26.3) | 38 (100) |
| Average size ±SD (mm) | 41.7± 16.0 | 26.5± 14.9 | 34± 5.7 | – | 39± 16.0 | 12.8± 6.3 | 19± 1.4 | – | 14± 6.1 | 32.4± 17.9 | |
| SD – standard deviation; ccRCC – clear cell renal cell carcinoma (RCC); pRCC – papillary RCC; chRCC – chromophobe RCC; AML – angiomyolipoma. | |||||||||||