Automated HER2 analysis predicts breast cancer outcome
MedWire News: Computerized assessment of human epidermal growth factor receptor (HER)2 immunohistochemistry (IHC) shows "excellent" agreement with visual assessment and is associated with outcome in patients with invasive ductal breast cancer, UK researchers report.
Zahra Mohammed (University of Glasgow, Scotland) and colleagues explain that HER2 status is an important indicator of breast cancer clinical outcome and response to therapy with trastuzumab.
HER2 status can be evaluated either by IHC for HER2 protein expression or by measuring HER2 gene amplification by fluorescence in situ hybridization (FISH). Current guidelines recommend IHC as a primary testing modality, with FISH reserved for ambiguous cases.
It has been suggested that computerized image analysis of HER2 IHC might result in more consistent interpretation than visual assessment, as it is less susceptible to intra- and interobsever variability.
To investigate, Mohammed and team compared visual with automated computerized image analysis of HER2 IHC status in tissue microarrays from 431 patients with primary operable invasive ductal breast cancer. They also assessed the ability of each method to predict patient outcome.
The researchers report that of 353 tumors scored using the visual method as 0 HER2, four (1%) cases were scored as 2+ and three (1%) cases as 3+ using the computerized method.
Of 23 tumors scored as 2+ HER2 using the visual method, five (22%) cases were scored as 0 and eight (35%) cases as 3+ using the automated method. And among 55 tumors scored as 3+ HER2 using the visual method, two (4%) cases were scored as 0 and three (6%) cases as 2+ with the computerized method.
This meant that the agreement between the visual and computerized methods was "highly significant," Mohammed and co-authors remark.
They also observed excellent correlation between HER2 FISH and both the visual and computerized image analyses of HER2 IHC.
During a median follow-up period of 165 months, 189 patients died, 93 as a result of their cancer.
Univariate survival analysis showed that visual and automated analysis of HER2 and HER2 FISH were each significantly associated with both recurrence-free and cancer-specific survival.
"The results of the present study show, to our knowledge for the first time, that HER2 status assessed by automated IHC is associated with recurrence-free and cancer-specific survival in patients with operable ductal breast cancer, and has similar predictive value to visual assessment and HER2 FISH," write Mohammed and co-authors in Histopathology.
"A good case can therefore be made for automated HER2 IHC in patients with primary operable invasive ductal breast cancer," they add.
The researchers caution, however, that automated IHC "will not remove the need for an expert human observer to determine that the images analyzed are of invasive carcinoma, representative of the lesion, and free from significant artifacts of microtomy or staining."
"Strict quality control will still be required during tissue handling, staining, image capture, and analysis," they conclude.
By Laura Cowen