Validating Currently Available Digital Measuring Tool for Melanoma
Gabor Hertz ; Kaiser Sacramento Medical Center;
Content:
Skin cancers account for the most common neoplasms seen in general clinical practice. Unlike the majority (basal cell carcinoma and squamous cell carcinoma), melanomas have a significant mortality rate, and have a reported incidence of 16-24/100,000.1 Treatment and prognosis of melanomas depends on the pathologic stage. Pathologic staging depends on measuring the Breslow tumor thickness. Traditionally this has been obtained by using an eyepiece reticle calibrated by stage micrometer. Pathology informatics/Digital imaging may provide a more reproducible measurement. To that end the study attempts to validate digital measuring tools available in an off-the-shelf product for use in staging melanomas.
Technology:
Olympus BX40 microscope, Olympus DP 71/DP 25 digital cameras using either FireWire or PCI connections, Olympus 0.01 mm stage micrometer and Olympus Micro- Suite 5 software (Melville, New York).
Design:
Validation study using a stage micrometer standard and retrospective review of 18 randomly selected cases from our files over the last two years. The slides were reviewed, digitally imaged, and measured by a single pathologist. Digital measurements were obtained at varying optical magnifications, and the mean measurement was compared to the reported clinical Breslow measurement.
Results:
Validating the digital measurements to the stage micrometer found a disagreement averaging 25 % (up to 0.3 mm) from the expected digital measurment. Reviewing the 18 cases from our files, a digital measurement was successfully obtained on all 18 cases with an average discrepancy of 0.12 mm between the analog and digitally obtained measurements.
Conclusion:
This validation study showed that the digitally obtained results varied up to 25% from the stage micrometer standard. This discrepancy may be a result of the stage micrometer, or technical problems with either the initial software installation, or a mismatch between the optical system and software. These factors need to be further elucidated prior to clinical use of this measuring tool. Interestingly, in the 18 retrospective clinical cases, there was tighter correlation between the analog and digitally obtained measurements (averaging 0.12 mm). This tighter correlation is probably due to multiple additional confounding factors inherent in pathologic staging that should be mitigated by the documenting abilities of digital imaging.
