06. Prognosis

Overview of Prognostication

  • Prognosis is a prediction for a particular future outcome (survival, functional status or quality of life) based on a person’s medical condition.
    • Prognostic information may help patients and surrogates make important health-related and life choices in the context of their medical condition(s) and expected course.  
  • Estimating prognosis is challenging – each patient is unique and it can be hard to generalize from the results of prognostic tools or clinical trials. There is inherent uncertainty in prognostication; it is important to acknowledge this with yourself and with your patients. 
  • Clinicians, patients, and surrogates often overestimate survival. Coupling clinical gestalt with objective data/tools can increase our ability to accurately prognosticate. 

Prognostic Tools for Patients with Cancer

  • Performance status (a measure of patient’s functional capacity) is a significant predictor of overall survival
    • Karnofsky Performance Score (KPS): measured from 0 to 100, with higher numbers indicating better function
      • KPS < 60: median survival <6 months
      • KPS < 50: median survoval 1 month
    • Eastern Cooperative Oncology Group (ECOG) Scale: measured from 0-5, with lower numbers indicating better function
      • ECOG > 2: median survival <6 months
    • Palliative Performance Scale (PPS): measured from 0 to 100, with higher numbers indicating better function
      • Correlates well with KPS and ECOG to estimate one-year survival in patients with advanced cancers
  • Prognosis based on complications of cancer 
    • Patients with Central Nervous System (CNS) Metastases
      • Stereotaxic radiosurgery (e.g. Gamma-Knife), modern targeted systemic therapies, and immunotherapies (e.g. anti-PD-1 therapies, CAR-T cells) have increased median survival in specific cancers with brain mets. Involve Radiation Oncology, Oncology, and Neurosurgery for multi-disciplinary discussion. 
      • Untreated CNS metastases from solid tumors: Median survival 1-2 months
      • CNS metastases treated with radiation: Median survival 3-5 months
      • Leptomeningeal metastases with treatment: Median survival 2-4 months
      • Prognostic factors: performance status, age at diagnosis, irreversible neurologic deficits at diagnosis, treatment history, and extent of extracranial metastatic disease.
    • Malignant Effusions: Median survival 3-12 months
      • Prognostic factors: tumor type (GI and lung cancers have worse prognosis compared to ovarian, breast, and lymphoma), tumor chemosensitivity, performance status
    • Malignant Ascites: Median survival 1-4 months
    • Hypercalcemia of Malignancy (suggests advanced cancer or heavy disease burden): Median survival 1-4 months

Prognostic Tools for Patients with Non-Cancer Conditions 

  • Geriatrics patients without a dominant terminal illness
    • Repository of prognostic indices located at eprognosis.ucsf.edu
  • Dementia 
    • Challenging to predict prognosis as patients have varying trajectories. 
    • Hospice eligibility (<6-month prognosis) for dementia is defined as profound memory deficits (unable to recognize family members), total functional dependence, speech limited to <6 words, incontinence, unable to ambulate without assistance, plus one dementia-related complication.
    • Common complications in the last year of life: infections (in particular PNA), eating difficulties, dyspnea (>5d/month), Stage 2+ pressure ulcers, unexplained weight loss.
  • Heart Failure
    • Challenging to prognosticate given patients’ fluctuating disease trajectory and inconsistent use of goal-directed medical therapy. 
    • NYHA Class IV (severe symptoms: unable to carry on any physical activity without discomfort; symptoms at rest) associated with 30-40% 1-year mortality.
  • COPD
    • The BODE index (BMI, exercise capacity, subjective dyspnea, and FEV1) in ambulatory COPD patients to estimate 1-3 year prognosis and risk of hospitalization. 
    • Patients with COPD who receive mechanical ventilation for any reason have 25-30% in-hospital mortality.
  • Cirrhosis
    • Higher MELD and MELD-Na scores are associated with increased three-month mortality risk in patients not undergoing liver transplant.
    • Child’s-Turcotte-Pugh (CTP) Classification predicts 1 to 2 year survival and likelihood of developing cirrhosis complications. 
    • Median survival in patients with decompensated cirrhosis is ~2 years. 
    • Hepatorenal syndrome portends poor prognosis: Type 1 associated with 8-10 week median survival even with treatment. Type 2 associated with 6-month median survival. 
  • End Stage Renal Disease
    • After stopping dialysis: Mean survival = 6-8 days (range 2-100 days, longer survival expected in patients with residual kidney function).
    • Charleston Comorbidities Index, performance status, and the “surprise” question ("Would I be surprised if this patient died within the next year?") can help identify ESRD patients that are at high risk for early mortality and facilitate discussions about GOC and end-of-life planning.

Tips for Communicating Prognosis with Patients and Families

  • Ask permission before discussing prognosis given the sensitive nature of this information.
  • See sections Breaking Bad News and Family Meetings and Communication About Serious Illness for more details. 
  • Use ranges of possible outcomes to acknowledge the inherent uncertainty in prognostic information. Examples include an estimated survival of hours to days, days to weeks, weeks to months, or months to years.
  • Acknowledge the uncertainty of prognosis and how challenging it may be for patients and surrogates to incorporate this information into decision-making.
    • When coupling prognosis with treatment options, consider using the “best case/worst-case/most likely case” framework (helpful video demonstration available online at https://www.youtube.com/watch?v=FnS3K44sbu0). 
  • Understand that discordance between patient/family and provider prognoses is common, even when the information has been delivered skillfully.
    • In ICU patients, physician-surrogate discordance on prognosis occurred in 53% of cases.
    • In addition to information from care providers, patients/families report using their prior experiences and personal, spiritual, emotional, and social beliefs to estimate prognosis.
    • If there is discordance, ask open-ended questions to identify the factors and sources of discordance to best help patients and caregivers process information and make decisions.

References

  1. Salpeter et al., J. Pall Med 2012, PMID: 22023378
  2. Gabani et al., Radiother Oncol 2018, PMID: 29960685
  3. Sloot et al., Cancer 2018, PMID: 29023643
  4. Gaspar et al., IJROBP 1997, PMID: 9128946
  5. Hyun et al., 2016 EJC 2016, PMID: 26841095
  6. Le Rhun, Galanis, Curr Opin Neurol 2016, PMID: 27661208
  7. Cheng, Perez-Soler, Lancet Oncol, 2018, PMID: 29304362
  8. Bibby et al., Eur Resp J, 2018, PMID: 30054348
  9. Goldner, J Oncol Practice, 2016, PMID: 27170690
  10. Eprognosis.ucsf.edu, by Alex Smith, Eric Widera, Sei Lee, and colleagues, UCSF.
  11. Mitchell et al., NEJM 2009, PMID: 19828530 (The CASACDE study)
  12. Celli et al., NEJM 2004, PMID: 14999112
  13. Ong et al., Chest 2005, PMID: 16354849
  14. Nevins, Epstein, Chest 2001, PMID: 11399713
  15. Seneff et al. JAMA, 1995, PMID: 7500534
  16. Wiesner et al., Gastroenterology, 2003, PMID: 12512033
  17. Hemmelgarn et al., AJKD, 2003, PMID: 12830464
  18. Moss et al., CJASN, 2008, PMID: 18596118
  19. White et al., JAMA, 2016, PMID: 27187301
  20. Hwang et al. Cancer Invest, 2004, PMD: 15581048
  21. Taylor et al. JAMA Surg, 2017, PMD: 28146230