Tumor Necrosis Commentary

Tumor necrosis is a parameter included in multiple published tumor grading systems in human and veterinary oncology. However, currently there are no standardized methods to assess tumor necrosis. We presume that this oversight exists because pathologists assume that their colleagues all assess necrosis the same way – indeed, necrosis is a basic pathologic feature we learn to recognize at the very beginning of our pathology training. Surprisingly though, through personal communication with other colleagues, we have found that we do not all assess necrosis the same way!  We must standardize our tumor necrosis assessments so we can improve our tumor grading systems.

The first, likely underappreciated, and arguably most important aspect of assessing tumor necrosis that must be standardized is the tumor trimming technique.  The importance of the trimming technique is likely underappreciated because pathologists typically are only involved in reading slides that have already been produced rather than in tissue trimming and processing.  Here we suspect the time-honored adage “out of sight, out of mind” likely rings true: when we receive our prepared slides, we assume that the trimming method has resulted in unbiased, representative sections of the tumor from which to make our assessments – but how valid an assumption is that?  Trimming technicians are often trained to avoid sampling areas that appear grossly necrotic. This unavoidably introduces bias into our histologic assessment of the amount of tumor necrosis! Regarding methods to produce representative sections of the tumor, which is especially important in heterogeneous tumors, Roccabianca P et al., in the recently published fascicle “Tumors of Soft Tissue” by the Davis Thompson Foundation, suggested to prepare one tissue block for each 2 cm of longest dimension of soft tissue tumors (STT). This recommendation could relatively easily be tested in STT and other tumor types, and we hope it is soon!  However, for most veterinary laboratories, one section per 2cm of tumor seems impractical for larger tumors due to the large number of processed sections which could result in prohibitively increased processing costs and increased reviewing time for pathologists. 

Once tumor trimming methods are standardized, it will be important to standardize how many (and which) of the available histologic sections pathologists should use to determine the amount of necrosis. Currently, one pathologist may assess necrosis based on the slide with the greatest amount of necrosis, another pathologist may assess necrosis based on the slide determined to have the average amount of necrosis of all reviewed slides, and a third pathologist may assess the amount of necrosis based on all available slides.  Which of these methods is “correct”?  We have no guidance in the literature to answer this, and the literature does not indicate which of these methods was/were used to develop our published tumor grading systems.  We suggest that the most representative method would be to use the amount of necrosis based on all available slides; however, this assumes that the available slides are representative of the tumor.  This is potentially a faulty assumption; therefore, determining standardized tumor trimming methods takes precedence.

Another aspect of standardizing tumor necrosis assessment worth discussing is determining appropriate cutoff values for the amount of necrosis.  Cutoff values not only must be shown to be prognostically useful (alone or in combination with other grading system parameters), but they must be reproducible with minimal inter- and intra-observer variance.  The 50% cutoff value for necrosis that is used in the grading system for canine STT would seem to be reproducible and relatively easy to assess both grossly (by technicians and inexperienced and experienced pathologists) and histologically (by inexperienced and experienced pathologists). The histological necrosis cutoff values of 1-<10% and 10%- <50% that are discussed in the VCGP Tumor Necrosis Guideline would be more difficult for pathologists to assess consistently, especially without computer measurement assistance. The authors of the Guideline proposed the following scoring system for necrosis: 0=no necrosis seen grossly or histologically; 1= minimal or no necrosis seen on gross exam and histologic necrosis estimated at < 50%; 2=multiple large areas of necrosis seen grossly and histologically > 50% necrosis.  This could be simplified to say, based on histologic only or gross and histologic assessment: 0 (no necrosis), 1 (<50% necrosis), and 2 (>50% necrosis); we recommend avoiding subjective terms such as “minimal” and “multiple large areas” when standardizing any methods because these terms introduce interobserver variability in interpretation. What is “minimal”? 1%? 5%? 9%?; what is “large areas”? 1 cm? 2 cm? 10 cm?

Furthermore, gross necrosis assessment opens somewhat of a Pandora’s box.  This parameter has never been assessed for validity, reproducibility, or usefulness in veterinary pathology although it is included in some human grading systems.  Current published veterinary tumor grading systems that use necrosis as a parameter do not indicate whether gross necrosis was used in the tumor necrosis assessment, and if so, whether it was confirmed histologically.  Moreover, there is no standard method of trimming tissue to assess for gross necrosis – should additional sections be examined grossly beyond what is processed for histology?  This would seem reasonable and would not increase cost or workload significantly.  For example, histologic samples may be trimmed first in the radial method or other appropriate method depending on tumor size, and then the remainder of the tumor can be grossly assessed using the “bologna” or “bread loaf” method with sections at 1-2cm intervals. The utility of gross necrosis as a parameter for tumor grading systems is in need of study: can a gross estimate of tumor necrosis be reproducible and adequately correlate to the histologic estimate?  We are concerned that this may not be possible.

Necrosis is currently used in multiple tumor grading systems in veterinary pathology, yet it is still fraught with issues and inconsistencies that need to be resolved! It is time to standardize.