SIU 2017: Prognosticating Factors: Clinical
Ultimately, we are all managers of health care risk – we seek to understand, predict and prevent future health care events!
Prognostic markers inform us regarding:
1) Who needs surgery?
2) Who should receive neoadjuvant or adjuvant therapy?
3) Who is eligible for clinical trials?
4) Who should undergo surveillance after treatment?
5) How to counsel patients
6) How patients perceive their disease
There are numerous prognostic markers in clear cell RCC (ccRCC). However, while many are proposed, clinical features (stage, grade, histology) remain the basis of most prognostic tools. In his review of the numerous risk models available for localized RCC, all of which were developed in a retrospective manner, the key features were inevitably TNM stage, grade, tumor size, performance status, presentation, age, gender and coagulative necrosis. Similarly, risk models for metastatic RCC are no different – primarily dependent on clinical features commonly reported.
- He does note that the predictive efficacy in localized RCC is superior to the metastatic RCC risk models
o C-index for localized risk models ~0.7-0.9
o C-index for mRCC risk models ~0.6
Full-spectrum “-omic” heterogeneity of RCC makes it unlikely that a single biomarker will be able to adequately discriminate outcomes better than current clinical histopathologic features. Therefore panels/profiles are more likely to be the future. But the difficulty is exceeding the discriminatory ability of current readily available data – stage, grade, histology!
Additional Clinical Prognostic Markers – none ready for prime-time yet, but have promise.
1) Radiomics
- Only approximately 5% of the data generated by a CT scan or MRI is used in the final rendering of the imaging – the remaining 95% has a significant information that to date has been untapped
- Radiomics is a young, nascent field that has the potential to significantly impact outcomes
- It utilizes image signal data, physical features, radiomic features and additional layers (genomics, etc)
- He provided some examples to help illustrate his point
- Sestamibi scan for discrimination of oncocytoma from RCC
- G250 PET imaging for ccRCC – primary tumor or metastatic disease
2) Tumor Complexity Metrics
- Tumor complexity, though previously described for operative planning and risk of complications, has also been shown to correlate with outcomes
- Nephrometry score, among others, represents an established tumor complexity measure
- Please refer to Dr. Uzzo’s prior talk on tumor complexity for further details on this topic (also on Urotoday.com from a Thursday session)
3) Pathologic Markers beyond TNM
- Beyond the standard clinicopathologic markers, we are all aware of other features known to be associated with outcomes – however, they have not yet been incorporated into risk profiles
- Examples:
- Variant histology presence – sarcomatoid, rhabdoid features
- Tumor necrosis
- Macroscopic vs. microscopic
- Likely more prognostic in setting of ccRCC than papillary
- In high-grade RCC, likely a poor prognostic indicator
- In low-grade RCC, it may actually be a good prognostic indicator
- Collecting system invasion
- Microvascular invasion
- Capsular penetration
4) Serum markers
- Many abstracts at each of the conferences focus on identifying clinical laboratory markers as prognostic or predictive
- Serum biomarkers – some are well-established, such as the following:
- MSKCC – used LDH, hemoglobin and corrected calcium to risk stratify
- Heng index – uses MSKCC but adds neutrophils and platelet levels
- Other potential markers being evaluated
- Elevated ESR, CRP
- Lymphopenia
- PDL-1 expression
- serum amino acid levels
- They all represent inflammation, immune dysfunction or metabolic dysfunction
Speaker: Robert Uzzo, United States
Written by: Thenappan Chandrasekar, MD, Clinical Fellow, University of Toronto, Twitter: @tchandra_uromd at the 37th Congress of Société Internationale d’Urologie - October 19-22, 2017- Lisbon, Portugal