Cultivating Prognosis In Non-Small Cell Lung Cancer

“It appears to me a most excellent thing for the physician to cultivate prognosis…”

Back in 400 B.C., when Hippocrates of Kos — father of clinical medicine — published his manuscript entitled “The Book of Prognostics,” no one could grasp — to its genuine extent, at least — the vanguard and insight of this endeavor towards shaping the prognostication process in modern medicine. Yet, this emphasis on prognosis when medicinal therapy was still in its infancy, was won its spurs over the centuries to become from that point forward a critical cornerstone of the medical practice.

The concept of forecasting, predicting, and prognosticating the course of a disease has certainly evolved over the years. Nevertheless, it is still bound to the terrain on which it occurs. Since patient uniqueness from a genotypic/phenotypic point of view was proven, the terms “personalized medicine,” “precision medicine,” “stratified medicine,” “theranostics,” to cite a few, have risen in usage. And given the latest breakthroughs in the understanding of the molecular basis of diseases (genomics), especially cancer (oncogenomics), this terminology seemed to fit perfectly within the lexical field of cancer management.

Usually, the more aggressive the disease is, the more burdensome the treatment must be. The reverse is equally true. Assessing how mild/aggressive the disease can be, however, a tough nut to crack. All we need is a yardstick of quality, a reliable tool to address the heterogeneity of prognosis.

In non-small cell lung cancer (NSCLC), for instance, the TNM system has made a giant leap towards stratifying patients according to the severity of their disease, with the purpose of smoothing the way for the most personalized therapy each patient should benefit from. Nonetheless, an overlap in survival rates (hence in prognosis) between stages remained within the TNM classification. This was an incentive for us to aspire to a more accurate staging tool, which would obviously include the TNM system, plus all relevant factors that could affect the outcome of the disease.

Since a comprehensive approach going through large numbers of prognostic variables might yield results that are hard to interpret, we believed that simple, easily-calculated parameters would suffice to sketch a more accurate prognosis in NSCLC patients. Conducting our work in a developing country such as Tunisia — where access to costly tools could be problematic — bolsters our choice.

This is a longitudinal survival study of a monocentric cohort of consecutive male patients with NSCLC. Data were collected using the patients’ case records, including: age at diagnosis, history of tobacco smoking, body mass index (BMI), performance status (PS) according to the Eastern Cooperative Oncology Group (ECOG) criteria, as well as clinical symptoms at presentation (cough, chest pain, hemoptysis, dyspnea, dysphonia, fever, lethargy, anorexia, and weight loss).

Biological data were provided by blood tests taken upon admission: hemoglobin level, white cell and platelet counts, and levels of C-reactive protein (CRP), calcium, lactate dehydrogenase (LDH), D-dimer, and alkaline phosphatase (ALP). Tumor-related factors were also collected: endoscopic site of lesion, histological type, TNM status (according to the AJCC 8th Edition Cancer Staging System for NSCLC), and type and number of metastatic sites. Data needed for calculating survival were collected prospectively and death was considered the outcome of interest.

The statistical analysis relied on the Kaplan–Meier method followed by Cox regression modeling. Two models were constructed, yielding 4 prognostic indices which were compared by their receiver operating characteristics (ROC) curves. The index with the best discriminative ability was picked and allowed a relevant risk stratification (5 risk groups in total). It included the following nine variables as prognostic factors: age, performance status, hemoglobin level, leucocyte count, calcium, lactate dehydrogenase, alkaline phosphatase levels, histological type, and TNM stage.

The value of the prognostic index ranged from 0 to 380. Each group was defined by an interval of 30 points, except for the last one (Cf. Table). Validation on the derivation cohort showed that our index impacted survival heavily, using simple parameters whose external validity has been proven in the literature.

These findings are described in the article entitled Development and validation of a prognostic index for survival in non-small cell lung cancer: Results from a Tunisian cohort study, recently published the journal Cancer Epidemiology. This work was conducted by Ghassen Soussi, Nissaf Ben Alaya, Nawel Chaouch, and Hajer Racil from Abderrahmen Mami Hospital, Ariana, Tunisia.