Tissue and Symptoms

Cancer is the second leading cause of death in the United States.  The possibility of cancer is often stressful and creates a feeling of uncertainty about the future.  From a pure medical viewpoint, a symptom is a malady the patient experiences and a sign is something that can be observed.  For example, a headache would be a symptom whereas a purple rash would be a sign.  If the purple rash caused itching, then the rash would be a sign and the itching (pruritus) is a symptom.  This series of articles will use the term symptom (i.e. as any uncomfortable feeling or complaint) from a patient’s viewpoint and not distinguish strictly between a sign or symptom.

There are several ways one can classify symptoms of a specific cancer, such as lung cancer.  For example, we could describe:

  • symptoms that suggest the presence of a cancer
  • symptoms that arise directly from the cancer
  • symptoms that imply the cancer has advanced in stage locally (or spread outside of its neighborhood, so to speak)
  • symptoms that imply the cancer has metastasized (or simply spread to other areas of the body)
  • symptoms related to treatment of a cancer
  • symptoms related to complications of a treatment for a cancer

In the above example (lung cancer) we could list symptoms based upon the above classification (note that there will be some overlap):

  • symptoms that suggest the presence of a cancer
    • Weight Loss, Cough, Pneumonia
  • symptoms that arise directly from the cancer
    • Hemoptysis (coughing up blood), Shortness of breath
  • symptoms that imply the cancer has advanced in stage locally (or spread outside of its neighborhood, so to speak)
    • Severe shortness of breath, Chest pain and/or pressure
  • symptoms that imply the cancer has metastasized (or simply spread to other areas of the body)
    • Swollen firm nodules in the armpits or near the collar bones, ulcer/bleeding from the skin of the chest, severe bone pain or bone fractures, severe headaches
  • symptoms related to treatment of a cancer
    • Nausea, Vomiting, Diarrhea
  • symptoms related to complications of a treatment for a cancer
    • Intestinal bleeding, severe infection / fever

From a practical viewpoint, most people would want to know what constitutes the early symptoms of a particular cancer in the hopes of detecting early so that it might be treated most effectively or even cured.  Alternatively, one might wonder if a particular symptom they feel might indicate they have a certain cancer.  The emphasis of subsequent articles will be to describe in detail the key symptoms that might point to a particular cancer based upon a specific organ (such as pancreatic cancer) or based upon a specific cancer type that can affect several regions of the body (such as non-hodgkin’s lymphoma).

Just how important are so-called key symptoms of various cancers?  The facts are that most cancer symptoms are very non-specific and most people with these symptoms DO NOT have cancer.  An important study by Shapley and colleagues appeared in the British Journal of General Practice in 2010.  These researchers pooled results from 12 different studies of several thousand older adults.  They questioned how many people with common symptoms of cancer that went to their doctor eventually were found to have a cancer.  These studies were evaluated to figure out the positive predictive value of a particular symptom (PPV).  The PPV is the percentage of people with a “positive test” who actually have the disease.  In other words, when a test says that you have the disease – how accurate is the test.  A very accurate test , such as an HIV blood test would have a very high PPV.  A test that looks for a common result that could be the product of many different things would have a low PPV.

When the older adult complained of rectal bleeding (a common symptom of colon cancer), they found about a 5% positive predictive value for this symptom.   In other words, rectal bleeding can occur for many reasons and for every 20 old adults that went to the doctor with rectal bleeding only 1 of them was found to have colon cancer.  The other 19 likely had causes such as hemorrhoids or diverticulosis of the colon.     This also means that for every 100 older adults with rectal bleeding, a physician would have to perform 100 colonoscopies to find 5 cancers.

Balancing the risk and cost of testing to investigate every symptom requires judgment from the physician and discussion with the patient.  Most symptoms of cancer that prompt an investigation to rule out cancer require a test that has either (or some combination of):

  • A significant cost
  • Some amount of risk due to the invasiveness
  • Exposure to a significant dose of radiation
  • Risk of intravenous contrast dye allergy or injury to the kidneys

No test is without risk; a hole in the colon is made unintentionally in about 1 of every 1000 colonoscopies.  This also means, in order to find 50 colon cancers a physician would run the risk of placing a hole by accident in one of the healthy colons resulting in possibly an unnecessary or emergency surgery.   This also does not address the cost of the testing and life years saved by treatment of a cancer.

Shapley and colleagues found similar results for other common symptoms such as bloody urine and bladder/kidney cancer, nodules of the prostate and prostate cancer, coughing up blood and lung cancer, breast lumps and breast cancer, and difficulty swallowing and esophageal cancer.  Most of these key symptoms averaged a little greater than 5% positive predictive value for the disease in question.

So what then should we do with a symptom that might indicate cancer?  Clearly the answer is not to ignore it because “it’s probably nothing.”  We can be reassured that the odds are it is not cancer, but the decision to undergo testing for further proof should be carefully discussed with your physician.  Our hope is that through these series of articles, the emphasis will remain on public awareness of key symptoms of cancer thus preventing a missed window of opportunity for cure.

REFERENCES: 

  1. Positive predictive values of ≥5% in primary care for cancer: systematic review.  Shapley M, Mansell G, Jordan JL,JordanKP.  Br J Gen Pract. 2010 Sep;60(578):e366-77. Review. PMID: 20849687
  2. Abeloff, M.D. (2008). Abeloff: Abeloff’s Clinical  Oncology, 4th ed. Chapter 10.Philadelphia,PA: Churchill Livingstone – Elsevier
  3. Goldman, L ;  Schafer A.I.  (2011). Goldman: Goldman’s Cecil Medicine, 24th ed. Chapter 183.New York,NY: Elsevier Saunders.
  4. This article was originally published on September 3, 2012 and last revision and update was 9/4/2015.