Table 4.

Recommendations for future studies of EMT markers in clinical cancer primary tumors and patient outcomes

  1. Standardize definitions of positive and negative marker expression status across studies of a given EMT marker measured using the same laboratory technique. To facilitate this for studies of protein expression using immunohistochemistry, score using computer-assisted image analysis to obtain continuous measures of marker staining (e.g., H scores, average intensities) that can be used to develop clinically useful cut points.

  2. Make study sample sizes as large as possible.

  3. For every EMT marker measured, present Kaplan–Meier survival curves stratified by marker expression status as well as Cox time-to-event modeling of the association between marker expression and patient time-to-mortality (or other outcomes).

  4. Develop consensus about what set of covariates should be included in multivariate models to produce valid estimates of the association between EMT marker expression in primary tumor cancer cells and patient time-to-mortality (or other outcomes).

  5. Measure EMT markers of interest as both RNA and protein in the same set of tumors to clarify which form of marker expression is more clinically useful.

  6. For studies using immunohistochemistry, report quantitative measures of inter-rater and intra-rater reliability for scoring images.

  7. To assess marker expression heterogeneity throughout a primary tumor, measure the marker in multiple cores taken from different parts of the tumor, including the invasive front, tumor center, and surface of the tumor away from the invasive front.