Kiatkowski, DJ et al.
Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 1, 2001
Reviewers: John Han-Chih Chang, MD and Kenneth Blank, MD
Source: Journal of Clinical Oncology 1998; volume 16 (number 7): pages 2468 ? 2477
Non-small cell lung cancer is one of the most deadly malignancies secondary to its high metastatic potential. The overall cure rate (5-year survival rate) is approximately 10 ? 15% if you include all patients with that malignancy. There is a subset of patients, namely stage I, who have a much better but guarded prognosis. By definition, stage I lesions must be localized without invasion into parietal pleura, chest wall or vertebrae. Primary tumors must be more than 2 cm from the carina and without metastases to lymph nodes. The 5-year survival rate after definitive surgery ranges from 60 ? 75% for these patients, but they only comprise 15 ? 25% of the all non-small cell lung cancer patients. Despite the "good" prognosis that being stage I conveys, 25% or more still succumb to their malignancy. Perhaps further adjuvant therapy is warranted for some patients. This article from Kiatkowski et al attempts to determine possible prognostic factors that may delineate those of the stage I patients who are at higher risk for recurrence.
This reportedly was a prospectively designed study to retrospectively develop a predictive model of patterns of recurrence in stage I non-small cell lung cancer. Patients from one institution (Brigham and Women's Hospital in Boston, MA) were evaluated. During a 9-year period from 1984 to 1992, 317 patients with stage I non-small cell lung cancer underwent resection. All patients had a mediastinoscopy or surgical sampling of the mediastinum and hilar lymph nodes. Seventy-three patients were excluded for various reasons (most for lack of follow-up data and more than one primary cancer). Thus, 244 patients were considered evaluable.
Seventy-one percent of the patients had a lobectomy, while the others had resections with a required 1.5 cm cancer-free margin. All patients had at least 3 years of follow-up. Median follow-up was 65 months. Cancer-free survival (CFS) was calculated from the time of surgery till any evidence of cancer recurrence or death (from cancer), while those that died without evidence of cancer were censored at the time of death. Failure-free survival (FFS) was calculated from the time of surgery to cancer recurrence and/or cancer death. Overall survival (OS) was calculated from the time of surgery to death.
Patient demographics, histopathological features and molecular factors were analyzed. These included: age, gender, extent of surgery, histology, adenocarcinoma subtype, differentiation, T1 versus T2, diameter, pleural invasion, lymphatic invasion, atelectasis, vascular invasion, plasma cell infiltration, tumor giant cells, mitotic index, p185neu, p53, rb, bcl-2, H-ras-p21, blood group A, K-ras mutation.
The 3 and 5-year CFS rates were 76% and 67%, respectively. Univariate analysis yielded significant decreases in CFS associated with age older than 60 years, wedge resection (as opposed to lobectomy or pneumonectomy), lymphatic invasion and World Health Organization (WHO) adenocarcinoma subtype: solid tumor with mucin production. FFS rates under univariate analysis had the same variables as significant with the minor exception ? the age cutoff was 50 instead of 60 years old. Under the same univariate evaluation, OS had the same significant factors as FFS rates with the addition of vascular invasion also contributing to a decrease. Again, under univariate analysis, evaluating molecular markers revealed that only immunostaining positive with p53 yielded a significant decrease in CFS, FFS and OS.
The single controllable poor prognostic factor detailed above is extent of resection. It has been well studied and documented that the greater the extent of resection, the better local control rate. Thus, a subset analysis of patients that received more than a wedge, partial or segmental resection would be helpful. Univariate analysis performed on patients who had a lobectomy or pneumonectomy (excluded 58 patients) resulted in lymphatic invasion, WHO adenocarcinoma subtype ? solid tumor with mucin production, lack of bcl-2 expression, and p53 immunostaining as poor predictors of outcome in CFS. FFS and OS had the same univariate factors with the addition of male gender and pleural invasion as significant negative influences.
Continuing to look at the patients with greater extent of surgical resection, more statistical manipulation with the Cox proportional hazards regression model revealed six independent and significant factors associated with decreased CFS. These were tumor 4 cm or greater, lymphatic invasion, adenocarcinoma solid tumor with mucin production, p53 expression, K-ras mutation, and lack of H-ras p21 expression. OS and FFS rates had similar factors in the Cox proportional hazards regression model with the substitution of bcl-2 production and male sex for K-ras mutation. OS in addition had pleural invasion as a significant negative factor.
What does this all mean?! Based on their findings, the authors feel that no patient should have anything less than a lobectomy, if they can tolerate the resection. They conclude that there are 6 significant prognostic factors affecting survival and/or recurrence in pathological stage I non-small cell lung cancer: adenocarcinoma solid type with mucin production, lymphatic invasion, tumor diameter, p53 expression, K-ras codon 12 mutation, and H-ras p21 lack of expression. The authors propose a new substaging of stage I non-small lung cancer. The data suggest a 5-year CFS of 87% for those with 2 or fewer negative prognostic factors (Ia) , 58% for 3 (Ib), and 21% for greater than 4 (Ic).
Again, what does this mean to patients, to physicians?! Well, it appears that we have another system to predict prognosis. It is dubious that anyone will start referring truly pathological stage I patients for adjuvant therapy based on this model. The study was well designed, but all it has provided is a prognostic model. The only way this could be put to good use is if this was confirmed on a larger scale trial of multi-institutions. Ideally, from there, a randomized trial of the Ic and perhaps the Ib patients could randomize to post-operative chemotherapy and radiation therapy versus observation post-operatively. Maybe then, if a benefit is elicited in adjuvant therapy, this current information is useful. For now, the data are overwhelming in favor of surgery alone for pathologically stage I non-small lung cancer versus surgery followed by adjuvant radiation therapy and/or chemotherapy.