Neural network‐based screening (NNS) in cervical cytology: No need for the light microscope?

Abstract

Neural network-based screening (NNS) of cervical smears can be performed as a so-called "hybrid screening method," in which parts of the cases are additionally studied by light microscope, and it can also be used as "pure" NNS, in which the cytological diagnosis is based only on the digital images, generated by the NNS system. A random enriched sample of 985 cases, in a previous study diagnosed by hybrid NNS, was drawn to be screened by pure NNS. This study population comprised 192 women with (pre)neoplasia of the cervix, and 793 negative cases. With pure NNS, more cases were recognized as severely abnormal; with hybrid NNS, more cases were cytologically diagnosed as low-grade. For a threshold value > or = HSIL (high-grade squamous intraepithelial lesions), the areas under the receiver operating characteristic (ROC) curves (AUC) were 81% (95% CI, 75-88%) for pure NNS vs. 78% (95% CI, 75-81%) for hybrid NNS. For low-grade squamous intraepithelial lesions (LSIL), the AUC was significantly higher for hybrid NNS (81%; 95% CI, 77-85%) than for pure NNS (75%; 95% CI, 70-80%). Pure NNS provides optimized prediction of HSIL cases or negative outcome. For the detection of LSIL, light microscopy has additional value.

Using classification and regression trees, liquid‐based cytology and nuclear morphometry for the discrimination of endometrial lesions

“…[4][5][6][7][8][9] More specially, concerning cytology, these techniques have been applied to various organs, among others stomach, [10][11][12] breast, [13][14][15][16] urinary system, [17][18][19] cervix, [20][21][22][23][24] and thyroid. [25][26][27][28] Classification and Regression Trees (CARTs) are very popular techniques, and have been often used in the past.…”

Diagnostic Parameters in Liquid-Based Cervical Cytology Using a Coagulant Suspension Fixative

“…12 All slides were of a good-enough quality to be scanned with the neural network scanner. 8 The mean time to scan the thin-layer slides with the scanner was 8 minutes, while it took 15 minutes for conventional slides. 11 The images of the Papspin ® slides had, with-out exception, a clear background, resulting in high contrast between the diagnostic elements.…”

“…After staining the slides were coverslipped with plastic film and scanned with a neural network scanner, formerly known as PAPNET. 8 …”

Clinical and diagnostic significance of blood in cervical smears

“…13 Since the introduction of the PAPNET testing system in the Leiden Laboratory, which uses neural network-based technology in combination with light microscopy, the balance between correctly diagnosed cases and "false-negatives" has been fully optimized. 8 This system is not commercially available in the United States. It is not unlikely that, in due course, an alternative, but comparable neural network based system will be available for use, not only within, but also outside, the United States.…”

“…7 Using this screening system, smears previously reported as negative can be reanalyzed for missed cancer cells or cancerous material or for additional information of background features such as blood. 8 In the context of bloody smears, a variety of studies have found that with progression of cervical dysplasia, mean vessel density increases in the underlying submucosal connective tissue of the cervix. Thus, the onset of angiogenesis appears to be an early event in malignancy-related changes of the cervix.…”