When Breath Becomes Air: The Screening That Would Not Have Found It

When Breath Becomes Air is a short memoir by a neurosurgeon who became a patient with metastatic lung cancer in the last year of his residency. Random House published it in January 2016, ten months after his death.

Stanford Medicine reported that Paul Kalanithi was diagnosed in May 2013 with stage IV non-small cell EGFR-positive lung cancer, at age 36, and that he had never smoked. He died on 9 March 2015 at 37.

Spoilers throughout. Kalanithi died before the book was finished. The epilogue is written by his wife, Lucy Kalanithi, and describes his last days. The publisher's own page says so. This page treats his death as the book's known ending, not a reveal.

It is not a lung cancer guide. It is a book about the moment a doctor moves from one side of a consultation to the other, and about what a person does with time that has stopped being open-ended. The publisher's page notes a foreword by Abraham Verghese and puts the book at 256 pages.

There is a short prologue, then two long parts, then the epilogue.

The prologue opens with Kalanithi looking at his own CT scan and reading it as a doctor reads any scan, then realising whose scan it is.

Part one, "In Perfect Health I Begin", is the pre-diagnosis life. An English literature degree, a turn toward medicine as the place where questions about meaning are answered in the body rather than on paper, then the grind of neurosurgical training: the hours, the operating room, the deaths, and the slow accumulation of competence.

Part two, "Cease Not till Death", is after the scan. Diagnosis. Treatment. Going back to operating. The decision to have a child. The decline. It ends mid-thought, because he stopped writing.

The epilogue by Lucy Kalanithi carries the story to the end, including the decision to stop treatment and the death itself. She is a physician too, and writes with the same clinical plainness.

It is strongest on the reversal.

Kalanithi knew what a stage IV scan means. He had delivered that news. Watching him try, and fail, to be his own doctor is the most honest thing in the book: he wants a survival number, his oncologist declines to give one, and he understands both why she declines and why he keeps asking.

He is also good on identity. Neurosurgery was not his job, it was who he was, and the question of whether to return to the operating room is not sentimental. He goes back. He operates well. Then he cannot.

And he is unusually clear about the arithmetic of a prognosis. What he wants is not a number but a shape: whether to plan in months or years, because the answer changes what he should do with the time. Many people describe that same problem and cannot name it as precisely.

There is a smaller strength worth naming. He is good on the ordinary indignities: the waiting, the scan-to-scan rhythm, the way a fortnight before a result swallows everything around it. He does not dramatise those. He records them, and then moves on, which is roughly what they feel like.

The clinical details are from 2013 to 2015 and are now dated. The drugs, the sequencing, the trial landscape and the expected duration of response in EGFR-mutated lung cancer have all moved since he was treated. Nothing about his regimen should be read as current practice.

It is also a very particular vantage point. He has medical fluency, colleagues who are also his doctors, an academic centre, and insurance. Almost nothing about cost, coverage, transport, translation or getting through a call centre appears in the book, because it did not have to.

The prose is literary and reaches for grand phrasing. Some readers find it consoling. Others find it too composed, especially in the middle of their own diagnosis, where nothing feels composed. Neither reaction is wrong.

And it has almost no practical content. There is nothing here to help a reader understand a pathology report, a treatment choice or a scan interval.

This is the gap the book unintentionally illustrates.

USPSTF recommends annual screening for lung cancer with low-dose computed tomography in adults aged 50 to 80 years who have a 20 pack-year smoking history and currently smoke or have quit within the past 15 years. That is a Grade B recommendation. Screening stops once a person has not smoked for 15 years, or develops a condition that substantially limits life expectancy or the ability to have curative lung surgery.

Kalanithi met none of those criteria. He was 36, not 50. He had never smoked, so he had no pack-years at all. There is no age or risk category in the United States that would have offered him a low-dose CT as screening.

That is not an oversight in the guideline. It reflects what screening can and cannot do. A test given to a whole population only helps if enough people in that population have the disease; below that point, false positives, follow-up procedures and radiation do more harm than the cancers found do good. NCI reports that 10 to 20 percent of lung cancers occur in people who never smoked, more often in women and at younger ages. That is a real number of people, and it is still too low a rate in far too large a population to screen.

It is worth being clear about what that means for an individual reader, because guideline language can sound like a judgement about who deserves attention. It is not. Eligibility for screening is a statement about a test's performance across a population, not about how seriously a person's symptoms should be taken. Someone outside the screening criteria who has a symptom is in exactly the same position as someone inside them: the symptom is the thing that gets investigated.

So for a never-smoker, and for anyone under 50, diagnosis comes through symptoms. NCI's list of possible cancer symptoms includes a cough or hoarseness that does not go away, unexplained weight changes, severe lasting fatigue, unexplained swelling or lumps, and back or chest pain, and its instruction is plain: if symptoms do not get better after a few weeks, see a doctor so problems can be diagnosed and treated as early as possible.

CDC's lung cancer risk factor page is worth reading alongside this, because it names things that are not smoking: indoor radon, which CDC says can cause lung cancer in people who have never smoked; workplace exposures including asbestos, arsenic, diesel exhaust, silica and chromium; HIV infection; previous chest radiation; and air pollution.

For the full eligibility picture see lung cancer screening, and for this specific group see lung cancer in people who never smoked.

Kalanithi's cancer was EGFR-positive, and that word is doing a lot of work.

NCI states that prognosis and treatment options in non-small cell lung cancer depend on whether the cancer has mutations in certain genes, such as the epidermal growth factor receptor gene, and that biomarker tests may be suggested to help predict response to targeted therapy. NCI lists osimertinib alongside erlotinib, gefitinib, afatinib and dacomitinib among the tyrosine kinase inhibitors used.

What has changed most is where osimertinib sits. NCI's drug page describes it as approved for EGFR-mutated non-small cell lung cancer in several settings: after surgery to help keep the cancer from coming back, in stage III disease that cannot be removed and has not progressed after platinum-based chemoradiation, as a first-line treatment on its own for metastatic disease, in combination with pemetrexed and platinum chemotherapy for metastatic disease, and after progression on another EGFR inhibitor.

None of that undoes the honest part. NCI says plainly that for stage IV non-small cell lung cancer, current treatments do not cure the cancer. A targeted therapy can hold disease for a long time and then stop working. That is the arc Kalanithi describes, and it is still the arc.

The reason this matters to a reader is narrow and concrete: the difference between his path and a chemotherapy-only path came from testing his tumor. See EGFR-mutated non-small cell lung cancer for what that testing involves. Kalanithi was 36 and had never smoked. No screening programme would have included him.

It suits readers who want to understand what a serious diagnosis does to a working life and an identity, and readers who are comfortable sitting with an ending that is already known. It suits clinicians, and people who love one.

It does not suit a reader who needs information. There is no treatment guidance, no explanation of staging, no help with decisions. It also may not suit someone recently diagnosed with lung cancer who is not ready to read a young doctor's death from the same disease, and there is nothing to be gained by pushing through that.

For the care he received in the last months, and what it is for, see palliative care.

• When Breath Becomes Air — publisher's page (Random House)
• Stanford Medicine — Stanford neurosurgeon, writer Paul Kalanithi dies at 37
• USPSTF — Lung Cancer: Screening
• NCI — Non-Small Cell Lung Cancer Treatment (PDQ) Patient Version
• NCI — Osimertinib
• NCI — Lung Cancer in People Who Never Smoked (press release, 2021)
• CDC — Risk Factors for Lung Cancer
• NCI — Symptoms of Cancer

This page discusses When Breath Becomes Air for education. It is not medical advice, and nothing here is a judgement of anyone's real medical care.