If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
In 2005, the European Association for Palliative Care made recommendations for prognostic markers in advanced cancer. Since then, prognostic tools have been developed, evolved, and validated. The aim of this systematic review was to examine the progress in the development and validation of prognostic tools.
Methods
Medline, Embase Classic and Embase were searched. Eligible studies met the following criteria: patients with incurable cancer, >18 years, original studies, population n ≥100, and published after 2003. Descriptive and quantitative statistical analyses were performed.
Results
Forty-nine studies were eligible, assessing seven prognostic tools across different care settings, primary cancer types, and statistically assessed survival prediction. The Palliative Performance Scale was the most studied (n = 21,082), comprising six parameters (six subjective), was externally validated, and predicted survival. The Palliative Prognostic Score composed of six parameters (four subjective and two objective), the Palliative Prognostic Index composed of nine parameters (nine subjective), and the Glasgow Prognostic Score composed of two parameters (two objective) and were all externally validated in more than 2000 patients with advanced cancer and predicted survival.
Conclusion
Various prognostic tools have been validated but vary in their complexity, subjectivity, and therefore clinical utility. The Glasgow Prognostic Score would seem the most favorable as it uses only two parameters (both objective) and has prognostic value complementary to the gold standard measure, which is performance status. Further studies comparing all proved prognostic markers in a single cohort of patients with advanced cancer are needed to determine the optimal prognostic tool.
Estimating prognosis is a fundamental component in the management of patients with advanced cancer for several reasons. First, accurate estimation of prognosis can help inform whether anticancer treatment is likely to be beneficial.
Third, it can help with end-of-life care planning, including place of care.
However, in patients with advanced cancer, the ceiling limit of the TNM classification system is often reached (i.e., M1) and as such is of limited value. As such, in the clinic, prognosis is based on various factors including stage of disease, performance status, previous clinical experience, and knowledge of cancer trajectories. However, the subjective nature of these may result in estimates of prognosis that are inaccurate, potentially misleading, and may result in anticancer therapies being given inappropriately.
In an attempt to improve prognostic accuracy, in 2005, the European Association of Palliative Care (EAPC) published recommendations on the use of prognostic markers in patients with advanced cancer.
Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care.
These recommendations were informed by eight studies examining different prognostic tools, which had been published in the preceding decade (1993–2003), and recommended a number of prognostic tools and their utilization. These tools were the Terminal Cancer Prognostic Score, the Palliative Performance Scale, the Palliative Prognostic Index, and the Palliative Prognostic Score.
Because these recommendations were made, a plethora of prognostic tools devised for use in patients with advanced cancer have been developed; however, to date they have not been presented together and comparison made. To this end, the aim of this systematic review was to examine and compare prognostic tools in patients with advanced cancer and make recommendations for their use.
Methods
The following databases were searched: Medline (2003–2015) and Embase Classic and Embase (2003–2015). The search focused on studies of prognostic tools in patients with advanced cancer regardless of the original primary tumor. The search terms are listed in Appendix I. A hand search of key journals and relevant citations was carried out. The date of the last literature search was April 30, 2015.
Eligibility Criteria
Eligible studies met the following inclusion criteria: population with advanced cancer (defined as an incurable cancer), original studies, study population n ≥100 and age ≥18 years, quantitative clinical and/or biomarkers were examined, a multivariate statistical model was described, the tool had been examined and validated in two or more independent data sets, published in English, published after 2003 (end date of original literature search), and full article was available.
Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care.
The primary outcome measurement examined was survival prediction (likelihood of death) based on the use of the prognostic tool in the specific patient population. Studies were excluded if a univariate survival analysis was described only, the tool was designed for use in one specific population with one specific cancer type (e.g., only patients with specific stage of lung cancer), or qualitative indices were used exclusively to predict survival.
Data Extraction and Analysis
The initial database search was undertaken and duplicates removed. Two authors (C. S. and K. M.) independently screened each study for eligibility based on the abstract and finally each full text article. From this, the necessary data for descriptive and quantitative analyses were extracted by C. S. and T. S., independently. These included the descriptors of the patient population, length of survival, and information regarding survival predictions. The analysis of each study was performed using standard quality assessment criteria which were then summarized for statistical analysis and comparison where possible.
Studies are presented according to the prognostic tool described. Where studies examined both populations with cancer and noncancer, only those populations with cancer were included in the analysis.
Results
The literature search process is shown in Figure 1. After abstract review, 179 articles were reviewed in full and this resulted in 49 studies fulfilling the eligibility criteria.
From the 49 eligible studies, seven different prognostic tools were identified. A summary of these is detailed in Table 1. The tools identified were the Palliative Prognostic Score (PaP, eight studies), Delirium-PaP (D-PaP, two studies), B12/C-Reactive Protein Index (BCI, one study), Prognosis in Palliative Care Study (PiPS, one study), Palliative Prognostic Index (PPI, eight studies), Palliative Performance Scale (PPS, 18 studies), and the Glasgow Prognostic Score (GPS, 10 studies).
A detailed description of these seven prognostic tools is given in Appendices II and III. These tools used a combination of clinical and/or biomarker parameters. The most common clinical parameters used were performance status, anorexia, and dyspnea. The most common biomarkers were C-reactive protein, white cell count, lymphocyte count, and albumin. The number of parameters used ranged from two (GPS, BCI) to 17 (PiPS B), and the mean number was seven. The largest single population studied for each of the prognostic tools is summarized in Table 2. Details of all studies included in this review are summarized in Supplementary Table 1.
Table 2Summary of Prognostic Tools—Largest Population Studied Per Tool
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Combination of initial Palliative Prognostic Index and score change provides a better prognostic value for terminally ill cancer patients: a six-year observational cohort study.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
To date, there have been eight studies (combined total n = 2694) examining the PaP in patients with advanced cancer. Patient cohorts were unselected but included patients with a variety of cancer diagnoses including cancer of the head and neck, lung, skin, breast, gastrointestinal tract, genitourinary tract, prostate, gynecologic, neuroendocrine, and hematologic tissue. The studies were from groups in Australia (one study), Italy (two studies), Brazil (one study), Japan (one study), Canada (two studies), and the U.S. (one study), thereby providing external validation of the tool. Two studies (n = 910) examined the D-Pap in patients with advanced cancer.
This included patients with cancers of the head and neck, lung, breast, gastrointestinal tract, and genitourinary tract. Both the PaP and D-PaP predict survival in patients with advanced cancer. The D-PaP tool has not been as extensively validated compared with the PaP; however, both perform similarly compared with each other.
The patient population included those with a diagnosis of cancer of the head and neck, lung, breast, gastrointestinal tract, genitourinary tract, prostate, gynecologic, neuroendocrine, and hematologic tissue. This study confirmed that an elevated BCI predicts poor survival.
The patients included those with diagnoses of gastrointestinal, lung, unknown primary, breast, urologic, gynecologic, central nervous system, hematologic, and head and neck cancers. This study reported that the area under the curve varied between 0.79 (PiPS A) and 0.86 (PiPS B) and suggested that PiPS is at least equal to and may be better than the clinician's predicted survival.
Eight studies (n = 5929) have examined the prognostic value of the PPI.
The patients included those with cancer of the head and neck, lung, breast, gastrointestinal tract, genitourinary tract, prostate, gynecologic, and hematologic tissue. The studies were based in Japan (three studies), Italy (one study), Taiwan (two studies), U.S. (one study), and Canada (one study). Recently, studies have examined a change in PPI scores, and this approach to researching the PPI appears more consistent, accurate, and clinically useful.
Eighteen studies (n = 21,082) have examined the PPS. The patients included those with diagnoses of cancer of the head and neck, lung, breast, gastrointestinal tract, genitourinary tract, prostate, gynecologic, neuroendocrine, and hematologic tissue. The studies were based in the U.S. (six studies), Spain (one study), Canada (eight studies), Italy (one study), Singapore (one study), and South Korea (one study), thereby providing external validation of the tool. Because of the numerous subgroups within the tool, earlier reports had stated it was not highly discriminating in the intermediate scores.
Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care.
Studies taking place after 2005 tackled this issue and focused on the significance of a 10% decrement in PPS score or poorer PPS scores. A strong ordering effect across the different PPS categories was demonstrated, with highly accurate scores for a PPS of 40% or less. Patients with PPS categories greater than 50% had lower hazard ratios than patients with lower PPS scores.
Ten studies (n = 5163) have examined the GPS. The patients included those with diagnoses of cancer of the head and neck, lung, skin, breast, gastrointestinal tract, genitourinary tract, prostate, gynecologic, neuroendocrine, and hematologic tissue. Eight studies were from groups based in the U.K., one study was from Japan, and one study examined data from an international biobank of patients, providing external validation of this tool.
A descriptive comparison of the individual clinical and biomarkers parameters included in the each of the prognostic tools is listed in Table 3. The number of markers ranges from two (GPS) to 17 (PiPS B). The PPS is composed of six parameters (six subjective), the PaP composed of six parameters (four subjective, two objective), the PPI composed of nine parameters (nine subjective), and the GPS composed of two parameters (two objective).
Table 3Clinical and Biomarkers Per Prognostic Tool
Parameter
Prognostic Tool
PaP
D-Pap
BCI
PiPS-A
PiPS-B
PPI
PPS
mGPS
GPS
Clinical marker
PS
x
x
x
x
x
x
CPS
x
x
x
Anorexia/decreased oral intake
x
x
x
x
x
x
Dyspnoea
x
x
x
x
Ambulation
x
Delirium
x
x
x
Activity
x
Evidence of disease
x
Edema
x
Global health
x
x
Breast cancer
x
Male genital organs
x
x
Distant metastases
x
x
Bone metastases
x
x
Liver metastases
x
Mental test score
x
x
Heart rate
x
x
Dysphagia
x
Weight loss—last month
x
Fatigue
x
Biomarkers
Lymphocyte count
x
x
x
White cell count
x
x
x
Neutrophil count
x
C-reactive protein
x
x
x
x
Albumin
x
x
x
Vitamin B12
x
Platelets
x
Urea
x
Alanine transaminase
x
Alkaline phosphatase
x
PaP = Palliative Prognostic Score; D-PaP = Delirium-PaP; BCI = B12/C-Reactive Protein Index; PiPS = Prognosis in Palliative Care Study; PPI = Palliative Prognostic Index; PPS = Palliative Performance Scale; GPS = Glasgow Prognostic Score; PS = performance status; CPS = clinician-predicted survival.
To date, there have been limited studies on the direct comparison of the prognostic value of the above tools. One study compared the performance of the PaP to the D-PaP, PPS, and PPI and concluded that the PaP showed superior accuracy and reproducibility.
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
The application of the Palliative Prognostic Index, charlson comorbidity index, and Glasgow Prognostic Score in predicting the life expectancy of patients with hematologic malignancies under palliative care.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
The application of the Palliative Prognostic Index, charlson comorbidity index, and Glasgow Prognostic Score in predicting the life expectancy of patients with hematologic malignancies under palliative care.
Since the European Association for Palliative Care recommendations for prognostic tools were published in 2005, there have been a number of prognostic tools developed, evolved, and validated.
Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care.
The PPS has been studied in the greatest number of patients, externally validated, and consistently predicts survival in patients with advanced cancer. Other prognostic tools of note that have been validated and consistently predict survival are the PaP, the PPI, and the GPS. In addition, the latter (based on the combination of C-reactive protein and albumin) has been extensively validated since the original review.
Most of the prognostic tools (PPS, PaP, and the PPI) depend largely on the assessment of functional status as a core component. Therefore, their use in routine practice has been sparse compared with Karnofsky Performance Score or the simplified Eastern Cooperative Oncology Group Performance Score.
In addition, the relatively complex scoring systems of these prognostic tools may have prejudiced their routine use, whereas the similarities but clear differences in these are confusing and make comparison challenging. Therefore, it would be important to rationalize these subjective assessments into a simpler scheme with as advocated by Harding et al.
The PRISMA Symposium 1: outcome tool use. Disharmony in European outcomes research for palliative and advanced disease care: too many tools in practice.
From the present review, it is also clear that many of the tools, such as PaP, PPI, PPS, and even performance status, are predominantly subjective and it could be argued that where possible, these should be made more objective. For example, one such way would be to examine if skeletal muscle mass is related to functional status and whether it can be a surrogate marker of physical function. This would seem plausible as skeletal muscle indices are increasingly recognized to have prognostic value.
Although various prognostic tools have been validated, they vary in their complexity, subjectivity, and therefore their clinical utility. The GPS would seem the most favorable as it uses only two parameters (both objective) and has prognostic value complementary to ECOG performance status, most commonly used assessment of patient physical function, in the oncology of advanced disease. Further studies, comparing all externally validated prognostic tools in a single cohort of patients with advanced cancer, are needed to determine the optimal prognostic tools.
The search strategy in the present review was comprehensive and included the main medical databases and a detailed search strategy (Appendix I). However, there were three notable studies not included in the review. Feliu et al.
reported the development and validation of a prognostic nomogram for terminally ill patients with cancer in almost 900 patients. However, it is of interest that the nomogram included the components ECOG-ps, lactate dehydrogenase, lymphocyte count, and albumin concentrations that have been used in other externally validated prognostic scores, such as PaP, that have been examined in the present review. The second study by Kim et al.
reported the external validation of PiPS A and PiPS B in 202 terminally ill patients with cancer. Finally, our search was limited to April 30, 2015. This excluded a large external validation study (n = 2426) of the modified PiPS A and PiPS B prognostic tools reported by Baba et al.
in May 2015. Nevertheless, the present review is therefore a step toward the viewpoint of Harding et al. that “it would be important to rationalize these subjective assessments into a simpler scheme with judicious selection and refinement of existing tools” (The PRISMA Symposium 1: outcome tool use. Disharmony in European outcomes research for palliative and advanced disease care: too many tools in practice).
The PRISMA Symposium 1: outcome tool use. Disharmony in European outcomes research for palliative and advanced disease care: too many tools in practice.
It is clear that with the exception of the GPS and contrary to the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) guidelines, hazard ratio and 95% CI have been reported inconsistently in the prognostic tools developed for use in patients with advanced cancer. This precluded meaningful meta-analysis in the present systematic review. Therefore, future research should directly compare these validated prognostic tools within all advanced cancer types using similar statistical approaches, in keeping with the REMARK guidelines.
The present systematic review updated a previous review published a decade ago. The majority of the prognostic tools examined had less than five independent reports of their prognostic value, and therefore, a meta-analysis of the validated prognostic tools was not meaningful and a formal estimate of bias was not carried out. However, the data from each article were presented in detail (Supplementary Table 1) enabling the reader to draw conclusions as to their quality and the likelihood of bias using standard criteria. As a result, the present systematic review is largely descriptive giving an update in the progress of prognostic tools in the field.
Several key aspects of prognostic tools remain elusive, and the present article was unable to address these due to paucity of primary data. To illustrate, it is not clear if certain tools have greater utility in specific tumor types and/or at certain points in the cancer journey. Furthermore, the potential role of these clinical tools in clinical practice is unclear as their usefulness in treatment stratification or place of care planning is unknown; both these are unlikely to be addressed unless such tools are incorporated into routine clinical practice.
It is also clear that another challenge is to implement the right tool at the right point in the patient's cancer journey. This is important as this can affect different aspects of care, for example, whether to treat with anticancer therapy, preferred place of death, etc. To date, the application of the right tool, at the right time, remains elusive and is likely to require a combination of mixed methodologies to achieve this.
Conclusion
Prognosis remains a central tenet of care in cancer and validated tools applied correctly may serve to improve patient care. Since the previous systematic review and recommendations, many prognostic tools that have been examined are not integrated into routine clinical care. It could be argued that the multitude of tools available may have actually confused clinicians as to the optimal tool for use. Furthermore, as performance status remains at the forefront of clinical decision making regarding prognosis, tools which build on this would seem preferable, for example, the GPS and ECOG-PS. To provide some clarity as to the optimal prognostic tool, studies are needed which compare all independent prognostic markers, in a single population. Such studies are eagerly awaited.
Disclosures and Acknowledgments
This work was supported by Medical Research Scotland (grant number 487/FRG).
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Cox proportional hazards: The hazard ratio associated with a one-unit increase in PPI score Survival of less than three weeks was predicted with a PPV of 86% and negative predictive value NPV of 76%.
Multivariate Cox proportional hazards model on predicting death within three weeks: Including gender, age, BMI, BT, systolic and diastolic blood pressures, PR, initial PPI, and ΔPPI.
Combination of initial Palliative Prognostic Index and score change provides a better prognostic value for terminally ill cancer patients: a six-year observational cohort study.
Multivariate Cox regression: Adjusting for age, gender, primary cancer origin, referring medical department, and the interval between the hospital admission and referral dates
Log-rank test: PPI Group A vs. Group B vs. Group C Cox proportional hazards regression analysis with backward selection: Incorporating age, sex, PaP, PPI, serum albumin, fat-free mass, unadjusted phase angle, handgrip strength, maximal inspiratory pressure, and standardized phase angle.
Glasgow prognostic score predicts prognosis for cancer patients in palliative settings: a subanalysis of the Japan-Prognostic Assessment Tools Validation (J-ProVal) study.
Univariate Cox proportional hazards modeling: The area under the receiver operating characteristic curve: To measure predictive accuracy in cancer patients and noncancer patients.
Proportional hazards regression model on overall survival: Including PPS, dyspnea, pain, fatigue, and agitated delirium. 10% Decrease in PPS results in HR of 1.65
Log-rank test on overall survival: Initial PPS groups Increasing HR with increasing PPS group Multivariable Cox proportional hazards model on overall survival: Including gender, diagnosis, site, and PPS. Increasing HR with increasing PPS group (PPS 20% [0.40] to PPS 70% [0.039])
Log-rank test on overall survival PPS Group A vs. Group B vs. Group C Cox proportional hazards model on overall survival: Including age, gender, race/ethnicity, and PPS.
Multivariate analysis with Cox proportional hazards model on overall survival: Including executed advanced directives, Medicare/Medicaid insurance, PPS, and gender.
Log-rank test on overall survival PPS groups compared Cox proportional hazards model on survival: Including age, gender, location, diagnosis category, and initial PPS. Increasing HR with PPS group (PPS 70 [0.056] – PPS 20 [0.54]).
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Cox proportional hazards model on overall survival: Including age, gender, diagnosis, initial PPS, and survival curve time in days, initial PaP, MMSE score, and presence/absence of delirium on initial consultation (PPS 90% [0.21] PPS 40% [0.45])
Multivariate Cox proportional hazards model on cancer-specific survival: Including lactate dehydrogenase, hemoglobin, calcium, white cell count, neutrophil count, albumin, and C-reactive protein.
A prospective longitudinal study of performance status, an inflammation-based score (GPS) and survival in patients with inoperable non-small-cell lung cancer.
Multivariate Cox proportional hazard model on overall survival: Including gender, histologic subtype, PS, the European Organization for the Research and Treatment of Cancer Prognostic Score, WBC count, neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, CRP, albumin, and mGPS.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
Glasgow prognostic score predicts prognosis for cancer patients in palliative settings: a subanalysis of the Japan-Prognostic Assessment Tools Validation (J-ProVal) study.
The PaP score was constructed by the Italian Multicentre and Study Group in Palliative Care and validated in patients with advanced incurable cancer using 30 day survival probability. The D-PaP (Delirium-PaP) is a modified version of the PaP, incorporating a delirium assessment that slightly improved the predictive accuracy of the PaP. The PaP and D-PaP are the only prognostic tools included in this review that use clinician-predicted survival (CPS) as one of their indices. The PaP has six parameters: four subjective (clinical) and two objective (biomarkers). The PaP and D-PaP both rely heavily on CPS, a subjective parameter that can add an extra 8.5 points to the total score (PaP maximum 17.5; D-PaP maximum 19.5). The other parameters (biomarkers and symptoms) contribute a maximum of 2.5 points making this tool heavily reliant on the clinician's expertise in prognostication (Table A1, Table A2).
A key component of the PaP is clinician-predicted survival. It has been argued that CPS is dependent on physicians having sufficient knowledge and experience to make assess this adequately. From the eligible studies, it was noted that oncologists' (i.e., nonpalliative care specialists) CPS was shown to be well calibrated but individual predictions imprecise. Using the CPS from nonspecialists still enabled, the PaP to predict the short-term survival (30 days) of patients with advanced cancer “reasonably well.” The inclusion of CPS, therefore, does not detract from the PaP score being a unique combination of physician's judgment, corrected and integrated with a series of other objective parameters, optimising the score. In spite of this, this tool is not used routinely. This may be because of its heavy reliance on CPS, and therefore, clinicians do not need to use a tool that weights their existing opinion heavily, and therefore, they could argue will not alter their survival estimate. The other components of the tool have been individually validated for their accuracy in estimating prognosis; however, the individual weighting of each parameter is not known because no study has compared every clinical and biomarker important in prognosis in advanced cancer.
B12/CRP Index
The BCI was developed by a group at the University of London, U.K., following the EAPC's recommendations in 2005. It was initially validated in patients with advanced incurable cancer admitted to an elderly care facility. It can estimate up to 90 day mortality. Of interest is that the BCI incorporates vitamin B12 levels as a marker of prognosis; the rationale for this is that increased levels are present in myeloproliferative disorders, hepatocellular carcinoma, and metastatic liver disease. It consists of two objective (biomarker) parameters, CRP and B12. However, vitamin B12 is not always analyzed routinely in patients and may explain the lack of further research into this tool (Table A3).
Prognosis in Palliative Care Study
The PiPS was developed in a UK population with locally advanced or metastatic cancer. There are two versions of the tool (PiPS A and PiPS B) and differ, in that PiPS B incorporates biomarkers when assessing survival. It predicts survival up to and greater than 55 days. The PiPS A has 13 subjective parameters, whereas the PiPS B has nine subjective and eight objective (biomarker) parameters. The PiPS, similar to other tools, relies on subjective parameters; however, in this case, they are orientated toward specific symptoms, signs, and disease burden, and many are suggested by the EAPC as individual prognostic factors. The relative weighting of each of the prognostic factors is not available in the public domain, instead the tool is accessed electronically and a score issued (Table A4).
Palliative Prognostic Index
The PPI was developed in Japan in 1999, in patients with advanced incurable cancer. It divides survival into three groups and estimates survival up to six weeks. Risk Group A (PPI score ≤4) has an estimated survival of more than six weeks. Risk Group B (PPI score 5) has an estimated survival of less than six weeks but greater than three weeks. Risk Group C (PPI score >6) has an estimated survival of less than three weeks. It consists of nine subjective parameters (the PPS, oral intake, edema, dyspnea at rest and delirium) and reports the presence or absence of signs and symptoms with similar weighting given to the different parameters. One of the parameters used is the PPS that is a prognostic tool in its own right. By incorporating the PPS into the PPI, more subjective parameters are incorporated, and while this may increase the prognostic accuracy, it may increases bias and the complexity and reduce clinical utility (Table A5).
Palliative Performance Scale
The PPS was validated in a palliative care population in Canada. It provides a percentage score based on subjective indices giving a survival estimate up to three months. Survival accuracy of intermediate scores has been noted to be variable. It consists of six subjective parameters. Many of these parameters are focused on aspects of performance status including ambulation, activity levels, and performance status itself. Performance status is the gold standard in assessing a patient's fitness; therefore, this tool is bias toward performance status in that synonyms of performance status are included as parameters (e.g., levels of ambulation, activity, and self-care). One of the other parameters is conscious level, which could have been objectified by incorporating the Glasgow Coma Scale (Table A6).
In conclusion, the PPS has been extensively studied in a large patient population with advanced cancer, including multiple cancer types. It has performed well in the majority of the studies looking at the tool individually, the only criticism being its better accuracy with lower PPS scores. It has also been compared several times with other prognostic tools with varying results and again demonstrates comparable accuracy to other tools with lower PPS scores. The components of this tool are heavily bias toward performance status and disease burden emphasizing the importance of these clinical markers in prognosis.
The Glasgow Prognostic Score
The GPS was originally developed in patients with non–small cell lung cancer and subsequently refined to the mGPS. The GPS combines CRP and albumin to give a score of 0, 1, or 2, with increasing score suggesting decreased survival: CRP <10 = 0; CRP ≥10 = 1 (albumin ≥35); and CRP >10 + albumin <35 = 2. It has been validated in individual cancer types in addition to large populations of patients with advanced incurable cancer.
The application of the Palliative Prognostic Index, charlson comorbidity index, and Glasgow Prognostic Score in predicting the life expectancy of patients with hematologic malignancies under palliative care.
The GPS is entirely objective as the information needed to calculate the score is based on biomarker results. The GPS has been developed since the EAPC's recommendations in 2005 and meets the requirements set that any prognostic tool is quick and easy to use, and its scoring system is very simple. The GPS is also able to predict survival accurately several months before death. It fulfills the EAPC's recommendations of being quick and easy to use, along with robust evidence of its accuracy (Table A7).
References
Anshushaug M.
Gynnild M.A.
Kaasa S.
et al.
Characterization of patients receiving palliative chemo- and radiotherapy during end of life at a regional cancer center in Norway.
Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care.
Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital.
Combination of initial Palliative Prognostic Index and score change provides a better prognostic value for terminally ill cancer patients: a six-year observational cohort study.
Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system.
The application of the Palliative Prognostic Index, charlson comorbidity index, and Glasgow Prognostic Score in predicting the life expectancy of patients with hematologic malignancies under palliative care.
The PRISMA Symposium 1: outcome tool use. Disharmony in European outcomes research for palliative and advanced disease care: too many tools in practice.
Glasgow prognostic score predicts prognosis for cancer patients in palliative settings: a subanalysis of the Japan-Prognostic Assessment Tools Validation (J-ProVal) study.
A prospective longitudinal study of performance status, an inflammation-based score (GPS) and survival in patients with inoperable non-small-cell lung cancer.
31231-3&id=gr1.jpg){kind=link}