Jump to content
  • entry
  • comments
  • views

Facts are stubborn things, but statistics are pliable



This is a blog about lung cancer survival statistics

One of the first instincts that kicks in for almost all of us who have been devastated with a (lung) cancer diagnosis is to ask how long have we left to live. Any web search for survival rates returns dismal statistics and it is hard not to panic or at least get quite demoralized. When I told friends about my recent incidental diagnosis, one of them responded with utter shock but added “oh listen, my sister was given 3 months but lived 4 years” (I believe her sister had breast cancer). I am sure she meant well and was basing her comment on what Googling lung cancer survival returns so was trying to make me feel better. I was stunned and did not know how to react to her comment about being given 3 months etc, after all my kid is 6 years old and as much as me living any day now is a blessing, 4 years is still a horrifically short time for a 49 year old mom to have left with her family.

With my initial panic and despair easing a bit now, a month after my lobectomy, and 2 months after that nodule was captured in an incidental scan, I started reading a bit more about lung cancer, including statistics. In this post, I will try to parse what some of the survival statistics and their differences. If you simply want a summary and no detail, you can stop reading here and jump to the summary!

I wondered about the subtle but important differences in survival rate definitions after reading a blog post by our forum’s Tom Galli about lung cancer statistics. Tom happened to mention that the cause of death was not factored into the commonly cited survival rates. That assertion very much surprised me since, at least for lung cancer, the average age of diagnosis is around 70 so there is likely to be quite a bit competing causes of death that I would have expected survival statistics to take into account: e.g., did the lung cancer patient also have heart disease and died of a heart attack? vascular disease and died of a stroke? This  prompted me to take a deeper look into what exactly is being reported as survival rates. Before getting into some of those, a disclaimer that all 5- and 10-year survival statistics for lung cancer are (by definition) out of date so do not reflect the reality of the last few years. This is very important since lung cancer treatment has dramatically changed/improved over the last few years. Therefore, as you read some of the material cited below, keep in mind that these numbers are OLD and thus may not be as relevant to us patients living with lung cancer in 2022.

Different cancer survival statistics exist [Ref. 1] . Some are best suited for research and policy while other are better for prognosis and clinical decision making. For simplicity, I will list 3 such measures:

  1. Overall survival: Ratio of (the number of lung cancer patients who remain alive 5 years after diagnosis) to (the number of lung cancer patients). Some also call this the all-cause survival, observed survival or crude survival.
  2. Relative survival: Ratio of (the overall survival of lung cancer patients) to (the overall survival for similar but cancer-free population, matched by age, sex and race).
  3. Cause-specific survival: Ratio of (the number of lung cancer patients who are not dead from their lung cancer 5 years after diagnosis) to (the number of lung cancer patients).

I had assumed 3 (cause-specific survival) is what is always being quoted in statistics but quickly learned that this is not the case. In fact some report 1 (overall survival) though most report 2 (relative survival) but almost no one reports 3 (cause-specific survival).

Relative survival can be overestimated, e.g., when a healthy screened effect exists, as was demonstrated in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, where participants in this screening trial had 30%–50% lower mortality rates for heart disease, injury, and kidney disease than expected.

Relative survival can also be underestimated, e.g., smoker lung cancer patients typically have lower life expectancy than the general population because they have higher risks of death from many other cancers, as well as from heart disease.

For prognosis, one would need a survival measure of the net effect of a lung cancer diagnosis, in other words, the chance of surviving assuming the lung cancer was the only possible cause of death. This is very hard to measure. Cause of death is often not accurately captured. Even if it was, the question is, will that matter much in terms of end result? Turns out the answer is not quite simple: It may matter for some cancers but not much for other cancers. A Nature paper studied this very question by including competing causes of death and reporting the 10-year cause-specific vs. relative survival for lung, breast, prostate, ovary, oesophagus and colorectal cancers [Ref. 2]. I highlight a few findings from this study:

  • Though relative survival is usually preferable to use, for some cancers (like lung and prostate cancer), relative survival is inaccurate.
  • For cancers of the lung, relative survival was lower than cause-specific survival. In fact, for all cancers except breast and prostate, relative survival was lower than cause-specific with particularly large differences observed for lung cancer, ovarian cancer and colorectal cancer.
  • For lung cancer, when the population hazard was inflated for smoking, survival estimates were increased (since smoking is a strong confounder for survival from all causes).
  • Error between the 2 survival measures vary for lung cancer, say for 45-54 year olds, cause-specific survival was 25% higher than relative survival. Note that this was not stratified for stage etc so hard to really use for personal prognosis.
  • Solutions to such error/discrepancy may include adding other demographic variables than age and sex, and to stratify relative survival calculation by cause of death.


  • Currently cited lung cancer survival numbers are outdated. They do not include the years that saw dramatic improvements in how lung cancer has been more effectively treated with new therapies.
  • Relative survival rate is commonly reported. These statistics do not take into account for actual cause of death. For lung cancer, this was shown to lower actual survival rate measures below what they really are.
  • With the above caveats, if you still want to check/consider/use available (old) survival numbers, make sure to at least look at statistics relevant to you, e.g., find the rates related to your own age, stage, sex, cancer sub-type, smoking history etc.
  • Lung cancer is still terrible and has much worse survival by any measure than say breast cancer but we are individuals so medians and means do not represent us as individual patients. Currently little is known as to where we are on the patient distribution curve, are we to the left or the right of the mean, by how much, we don't know yet.

I personally found this SEER Explorer App useful (SEER is the Surveillance, Epidemiology, and End Results Program, which provides information on cancer statistics) [Ref. 3]. You can look up some stratified numbers, including 5-year survival and conditional survival (conditioned on the patient having already survived 0, 1, 3, or 5 years since the cancer diagnosis). Remember, even there, the numbers are old (covering years 2012-2018) and some criteria is missing, say smoking history.

Tom ends his messages with: stay the course.

I am ending my blog with: don’t live as a statistic (admittedly, I am still trying to abide by that).


[Ref. 1] . https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4829054/

[Ref. 2]. https://www.nature.com/articles/s41416-020-0739-4

[Ref. 3]. https://seer.cancer.gov/statistics-network/explorer/application.html?site=612&data_type=4&graph_type=5&compareBy=age_range&chk_age_range_122=122&series=9&sex=1&race=2&stage=105&advopt_precision=1&advopt_show_ci=on&advopt_display=2 


Recommended Comments


Statistics is a complicated discipline understood by few of us these days. Obviously, you've been trained and understand. Thank you for this comprehensive explanation to our community.

Indeed, don't live as a statistic. Numbers are precise; they mean one thing. Statistics use numbers to add veracity to uncertainty. The result is misunderstanding and despair. Aim for life enjoyed every day. Funny, I've never seen a statistical projection for joy!

Stay the course.


Link to comment

Great and informative blog, Lily!

I don't ever look at statistics because I just try to do the best I can each day. I don't believe that the statistics apply to me.

We are all on our own individual journey and no one can tell you how long you may have, etc.

I continue to hope for new treatments for all of us!

Congratulations on your blog. I also wrote one in case you want to read it

Mine is called, The Roscopal Effect.




Link to comment

Thank you for sharing, I’m happy to say that I have never fit and use the statistics for lung cancer and have never really tried to focus on them. I am now at 7 1/2 with Stage IV and despite starting out with a very grim prognosis I am doing well and have now been off treatment for almost 6 years. I do remember looking briefly at the statistics when I was first diagnosed and it was depressing and overwhelming so from that point on I was determined to beat them !  🤍

Link to comment
Add a comment...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...

Important Information

By using this site, you agree to our Terms of Use.