What They May Mean Alone or Together

This article discusses the importance of various cancer symptoms and how their meaning may change either alone or in concert.  Basic concepts about cancer symptom statistics are introduced also.  A key principle in interpreting various symptoms that may or may not be related to cancer, is understanding the setting in which they occur.  A quite ridiculous example would be the complaint of feeling cold in a man that goes jogging in just shorts during a winter snowstorm.  The “symptom” of feeling cold is easily explainable given the setting it occurs in.  A middle aged woman with some swelling of the front of the neck and a symptom of feeling cold all the time is quite another story.  This symptom, given the profile of the person it has occurred in and the swelling of the neck, is likely due to an underactive thyroid gland and would require testing of the thyroid hormone levels to prove this.

In the same way, symptoms that might be related to a cancer can be assigned a relative value of importance or level of concern based upon the setting they occur.  For example, consider two 55 year old men each with a new nagging cough that does not seem to go away.  If the first man lives in the southwest, works outdoors, and seems to get a mild cold every time this year – then this could be possibly explained by his environment and seasonal allergies.  If the second man has smoked two packs of cigarettes a day since he was 16 and had a brother die of lung cancer at the age of 58 – then this symptom would require careful evaluation for a possible lung cancer.  Both or neither of these men could have lung cancer, but the setting in which it occurs and the risk factors of each man raises or lowers the level of suspicion.  In many ways a physician that evaluates such complaints has to have the mind of a detective and consider all the possibilities before jumping to a conclusion.

In addition to the setting and risk factors that a symptom occurs; the severity, duration and associated other symptoms may modify the importance of a particular symptom.  Considering the above example, if the first man also says he has the symptoms of a 25 pound weight loss over the past 2 months and both of his parents smoked in the home growing up then this may raise the index of suspicion for lung cancer.

Taken alone, most symptoms are fairly non-specific for a certain cancer.  Consider the following data:

  • About 5% of women with bloody nipple discharge will have breast cancer
  • About 10-12% of palpable breast lumps will prove to be breast cancer
  • Almost 15% of women with a palpable lump and a normal mammogram will have breast cancer
  • Over 50% of women with a breast cancer have a palpable lump
  • A mammogram can detect about 80% of breast cancers
  • Over 30% of breast cancers have no symptoms other than abnormalities seen on a mammogram

These statistics are very general and would vary for different types of breast cancer.  They would also vary for different women with greater or higher risk factors for breast cancer; such as having several close relatives with early onset breast cancer.  Do these facts sound confusing?  That’s because they ARE confusing.  Laundry lists of statistics about cancer symptoms are confusing both to the general public wanting to understand their symptoms and to physicians intending to provide the best care.

Knowing that out of 100 women with bloody nipple discharge only 5 will have breast cancer is not terribly reassuring to a women who suddenly notices bloody nipple discharge that does not seem to go away.  Statistics have several key functions in medicine, but reassurance is generally not one of them.  A woman with bloody nipple discharge that has not gone away (who invariably knows there is a small chance that she may have breast cancer) will want some type of reassurance in the form of an expert medical opinion that she does not have breast cancer.

Consider the statistics about finding small amounts of blood in the stool.  Many colon cancers cause loss of small or large amounts of blood in the stool so these findings would be concerning for the possibility of colon cancer.  Finding very small amounts of blood in the stool is about 30-60% sensitive for colon cancer, depending upon which test is used to measure it and how many times it is measured.  These same findings are about 70-90% specific for colon cancer.

Sensitivity is the probability of having a disease if the test says you have it.  In other words, sensitivity is the likelihood of detecting a disease if you actually have it.  Specificity is the probability of a normal test being normal in someone who does not have the disease in question.  In other words, specificity is the likelihood of a normal test occurring in a healthy person.  While colon cancers do tend to cause bleeding they do not bleed continuously and if they bleed the blood will not be found or seen uniformly in all stool that is passed.  Furthermore, there are perhaps a hundred possible causes of blood in the stool that are not from cancer.  Other common causes of blood in the stool include angiodysplasia, polyps, diverticulosis, irritable bowel syndromes, ulcers of the colon, meckel’s diverticulum, stomach ulcers, etc.  The tests that find blood in the stool might also react to some food that is eaten or pills taken such as iron pills.  If this were to happen the test would say it is positive (there is blood in the stool) when really there is not – this is known as a false positive.  If the test failed to detect blood in the stool that is actually there, this would be a false negative.

These same statistical principles apply to the interpretation of most symptoms related to cancer.  How should anyone make sense of it all?

Understanding what cancers occur commonly and what the commonest symptoms that they might cause will prevent a serious symptom such as rectal bleeding from being dismissed as “hemorrhoids.

From a physician’s point of view the keys to properly understanding and interpreting symptoms or complaints that may represent cancer are to

1) have a deep knowledge of normal human anatomy and physiology,

2) understand what cancers occur commonly and thus what is most likely to be seen,

3) at least a thumbnail knowledge of rarer types of cancer that although rare must be considered,

4) understanding an individual person’s unique risk factors that make a particular cancer more or less likely to be present, and

5) judgment and experience in dealing with common cancers.  These factors emphasize the importance of finding a competent physician that you can establish good communication with in order to meet all of these criteria.


  1. Fatigue, dyspnea, and cough comprise a persistent symptom cluster up to five years after diagnosis with lung cancer. Cheville AL, Novotny PJ, Sloan JA, Basford JR, Wampfler JA, Garces YI, Jatoi A, Yang P J Pain Symptom Manage. 2011 Aug;42(2):202-12. Epub 2011 Mar 12.  PMID: 21398090
  2. Detecting lung cancer as a cause of hemoptysis in patients with a normal chest radiograph: bronchoscopy vs CT. Colice GL. Chest. 1997 Apr;111(4):877-84. PMID: 9106564
  3. Mandel JS, Bond JH, Bradley M, et al. Sensitivity, specificity and positive predictivity of the Hemoccult test in screening for colorectal cancers. Gastroenterology. 1989;97:597-600.
  4. This article was originally published on September 3, 2012 and last revision and update was 9/4/2015.