The Goldilocks principle applied to number of parameters used in a grading system; and tumor marker vs grading system

According to Avallone et al (2021), a grading system is “a method to quantify the putative clinical aggressiveness of a neoplasm based on specific histological features.” In general, the more neoplastic tissue resembles normal tissue, the less aggressive it is, and the more abnormal neoplastic tissue looks, the more aggressive it is. However, some tumors are exceptions to this principle, and there are many histologic features that can be assessed and correlated with specific clinical outcomes to develop the most accurate grading system for each tumor type. 

Some histologic parameters may be more prognostic than others for a given tumor type. Clearly, those with the most prognostic significance are the ones that should be considered for development of the optimal grading system, but how many criteria should be included in a grading system? Some systems in the literature use only one (Valli 2011, Valli 2006, Heller 2005), while others use more (Kiupel 2011, Sanders 2019). Although using the most prognostically significant histologic parameter may provide useful prognostic information, given biological variability, there is no single test that is prognostically reliable for each individual tumor. For example, mitotic count (MC) has been shown to be of prognostic value as a solitary tumor marker for canine cutaneous mast cell tumors. A MC cut-off of 5 in one study (Romansik 2007) and a stratification of 3 groups of MC, with cut-offs at 1 and 7, in another (Elston 2009) were found to be prognostically significant. However, a third study reported that not all high-grade tumors identified by the multi-parameter 2-tier grading system were identified solely by a MC of 7 or greater (Kiupel 2011). While the prognostic significance of a MC cut-off of 5 has not been directly compared to the 2-tier system (and should be), addition of other prognostically significant microscopic features is likely to improve the prognostic significance of the system and be more meaningful for clinicians/oncologists. On the other hand, the addition of excessive numbers of parameters, some of which may add very little prognostic significance, makes the system too labor intensive for routine use with little benefit.

The Goldilock prinicple, i.e., the concept of “just the right amount”, should be applied to the number of histologic parameters used in grading system development, balancing the prognostic significance of the individual parameters with ease of use and observer concordance of the system.  The prognostic advantage of adding parameters to a grading system must be documented in studies using the same outcome endpoints.