GRACE :: Lung Cancer


Dr West

Tumor Size as a Critical Factor in Weighing Value of Adjuvant Chemo

download as a pdf file Download PDF of this page

While there are good reasons to not pursue chemo after surgery for stage I NSCLC, there are several factors that argue at least for strong consideration of adjuvant chemotherapy for higher risk patients. Because stage IB generally has a less favorable prognosis than stage IA, it’s not suprising that the debate about which patients should or should not be receiving post-op chemo has centered more on the stage IB population, which have much more commonly been included in trials testing the value of adjuvant chemotherapy. Now we’ll focus on why tumor size has emerged as probably the most important factor in borderline cases where we might consider adjuvant chemotherapy but aren’t definitely convinced.

In my last post I enumerated reasons to not pursue adjuvant chemo for stage I patients, in whom the separation between T1 cancers and T2 cancers is usually the size, with a cutoff of greater or less than 3 cm (involvement if the pleural lining on the outer edge of the lung can also make a smaller tumor a T2 lesion). In that post, I described several positive adjuvant chemotherapy trials that did include stage IB patients, including the IALT trial (abstract here), Canadian BR.10 trial (abstract here), and the ANITA trial (abstract here), but in all of these cases, the stage I patients showed little or no benefit while the benefits were much greater for stage II and IIIA patients. And then there is the very important CALGB 9633 trial (preliminary 2004 abstract here, revised 2006 abstract here) that randomized over 300 patients, all with resected stage IB NSCLC, to 4 cycles of carbo/taxol or observation alone. As described in detail in my prior post, the preliminary results of the trial were positive overall, with a 12% improvement in overall survival at four years in recipients of carbo/taxol, but the curves came together and were not significantly different for overall survival (but were better for progresison-free survival) for those receiving adjuvant chemotherapy. But at the time that the revised negative results for the overall study were presented, Dr. Strauss presented an analysis of his trial outcomes that showed that the stage IB patients who had cancers of 4 cm or larger had a significant benefit from chemotherapy, while there was no benefit from the patients with tumors that were less than 4 cm:

CALGB 9633 OS in 4 cm and more

(Click to enlarge images)

CALGB 9633 tumors less than 4 cm

Interestingly, although this hasn’t been published yet to my knowledge, my understanding is that the BR.10 trial that showed no separation of the curves for stage IB cancers actually shows the same result, with better results for chemo recipients in the patients who had tumors of 4 cm or greater.

Intuitively, this makes sense, since the risks of chemo are the same whether you have little to gain or a lot to gain from more treatment, and we also know that the risks of recurrence and death increase as the size of the tumor increases (abstract here):

Stage I NSCLC and tumor size

While most of us might have presumed this, it’s still helpful to see the actual data from lung cancer cases bear this out. And because of these, many experts recommend post-operative chemotherapy very rarely for people with stage I NSCLC tumors smaller than 4 cm but are much more likely to recommend it for a patient with a larger cancer.

There are a few additional factors that are worth considering for patients and oncologists weighing the pros and cons of adjuvant chemotherapy, and I’ll cover some of those extras soon. But tumor size has thus far been the most dominant factor in question of whether to pursue chemo in the stage I setting.

3 Responses to Tumor Size as a Critical Factor in Weighing Value of Adjuvant Chemo

  • dadawg001 says:

    Dr. West,

    It’s been about 14 years since I took a probability and stistics course. Can you perhaps make a brief post on the various statistical mumbo jumbo and maybe make it a sticky notes post so that we can easily refer to it?

    I can’t believe that I am the only one that has trouble getting a firm grasp on the numbers from these studies. This one in particular since the curves are so tight, the statistics are necessary to understand what results are statistically signficant and what is probably not.

    Maybe include an easy to follow discription of HR, CI, P, etc. Maybe something also about the significance of 5 year survival too (what is so magical about 5 year survival, why isn’t 7 or 8 years used?)

    Thanks for all the great work that you do. I hope that your wife and children appreciate how much you are helping so many people and how special you are. My wife nags at me when I have to work late, but I’m not in the business of saving people’s lives!



  • Dr West
    Dr. West says:


    Very good suggestion. Some of this is in Cancer 101, which you can find in the Subject Archives. Many of these I’ve mentioned in posts here and there, but I’d bet a HUGE proportion of the people using OncTalk are dropping in just occasionally and don’t catch these passing references. The stats should perhaps be part of the Core Concepts, which will be accessible from every part of the website.

    5-year survival — for cancers that aren’t very indolent, if you are alive at 5 years, you’re likely cured and not going to be affected by the same cancer in the future. This applies to most but not all lung cancers, but many other cancers are slower moving, and it’s just a milestone that isn’t clearly more important than 7 or 10 or more years. In breast cancer or prostate cancer, 10 year survival is another important endpoint.

    HR stands for hazard ratio, which is basically the change from what would be expected from a standard treatment. It’s measured in relation to 1, and below 1 means the result is better than standard, while above 1 means the investigational result is less effective than standard treatment. A hazard ratio of 0.60 is 40% better than standard, and 1.40 is 40% worse.

    CI is the confidence interval, which is the range of what would be considered possible where the hazard ratio actually is. For instance, if the HR is 0.60, the confidence interval might be 0.33 to 0.93, which means that there’s a 95% chance it’s between 0.33 and 0.93. If the CI includes 1, it’s not considered statistically significant, and there’s a 5% or greater chance that there’s no significant difference between the investigational approach and the standard treatment.

    And the P value is the likelihood that the difference between the investigational approach and the standard, control arm is due to chance alone. It’s also expressed as a decimal and is usually considered to be significant if the number is 0.05 or less, in other words that the likelihood of the difference is 5% or less.

    All of this should be readily accessible to members, because these terms come up in the vast majority of posts.

    -Dr. West

  • dadawg001 says:

    Excellent post, thanks Dr. West. That clears things up quite a bit.

    Now I’m ready to play mad scientist!


Leave a Reply

Ask Us, Q&A
Lung/Thoracic Cancer Expert Content


download as a pdf file Download PDF of this page

GRACE Cancer Video Library - Lung Cancer Videos




2015 Acquired Resistance in Lung Cancer Patient Forum Videos

download as a pdf file Download PDF of this page

Join the GRACE Faculty

Breast Cancer Blog
Pancreatic Cancer Blog
Kidney Cancer Blog
Bladder Cancer Blog
Head/Neck Cancer Blog
download as a pdf file Download PDF of this page

Subscribe to the GRACEcast Podcast on iTunes

download as a pdf file Download PDF of this page

Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon

Subscribe to
   (Free Newsletter)

Other Resources

download as a pdf file Download PDF of this page

Biomedical Learning Institute