Why two radiologists read the same scan differently

Because a scan report is a measurement and a judgement, not a photograph. Two trained radiologists can select different target lesions, place the calliper on slightly different edges of the same lesion, or compare against a different earlier scan. Radiologists call this inter-observer variation. It is expected, documented, and usually smaller than it sounds.

Measurement is the first source of difference. Under RECIST 1.1, the reader chooses up to five target lesions, no more than two per organ, and measures the longest diameter of each. A tumour edge is rarely a clean line. Two millimetres of judgement on a twenty-millimetre lesion is a ten per cent swing, and response categories are decided on percentages.

The comparison scan is the second. Response is never read from one study alone. One radiologist may compare against the baseline scan taken before treatment started. Another may compare against the smallest measurement you have ever recorded, called the nadir. The same current images then produce two honest, different-sounding conclusions.

Immunotherapy adds a third layer that chemotherapy does not. Immune activation can enlarge a lymph node, light up a lung shadow, or make a tumour look temporarily bigger because immune cells have flooded into it. One reader may describe reactive, inflammatory-looking nodes. Another may report new nodes. Both are describing the same pixels with different degrees of caution.

Response-assessment PET-CT for CION patients is coordinated at partner imaging centres, which perform the scan and issue the report — CION does not own those imaging facilities. This page explains general reporting terminology only. It does not interpret your individual scan or report; that is a conversation for your treating oncologist.

Categories and thresholds shown are the standard RECIST 1.1 and iRECIST definitions referenced in NCCN, ASCO and ESMO guidance, indicative as of August 2026. Some reports use PERCIST or a purely descriptive style instead. If your report does not state which criteria were applied, it is reasonable to ask.

• Different target lesions — each reader may pick a different set of up to five lesions to track, so the two sums are not measuring the same thing.
• Different edges — tumour margins blur into normal tissue, and two callipers placed a couple of millimetres apart change the percentage.
• Different comparison scan — baseline, the previous scan, or the nadir are three different reference points, and each gives a different percentage change.
• Different scan protocol — contrast timing, slice thickness and whether the study was a plain CT or a PET-CT all change what is visible and measurable.
• Different machine or centre — scanners differ in resolution, and analog and digital PET-CT do not resolve small lesions identically.
• Immune-related findings — inflammation, sarcoid-like nodes and immune infiltration can look like disease to one reader and like reaction to another.
• Different criteria — a report written to RECIST 1.1 and one written to iRECIST can label the same images differently, entirely correctly.

Neither one automatically. A radiologist reports what the images show. Your treating oncologist decides what the images mean for your treatment, reading them against your symptoms, weight, appetite, examination and blood tests. Those are two different jobs. The report that matters is the one your oncologist has actually seen alongside you.

Give your oncologist both reports and the images. A second reader cannot re-measure anything from a PDF. Ask each imaging centre for the DICOM files, on a CD or a secure download link, along with the earlier comparison studies. With the images in hand, an oncologist can see whether the two readers measured the same lesions at all — which is very often where the disagreement ends.

Then let a group read it, not a fourth stranger. Every CION patient's imaging is discussed at a tumour board, where medical, surgical and radiation oncologists look at the same study together. A group reading is a far more useful response to two conflicting reports than a race between centres, and it does not cost you another scan.

If a report seems to contradict how you actually feel, say so out loud at your consultation. Pseudoprogression — a tumour looking temporarily larger while treatment is working — is one recognised reason a scan and a patient can appear to disagree, and it is exactly why iRECIST asks for a confirmation scan before calling progression.

1. 1

   Ask whether the disagreement changes anything

   Put the question to your oncologist directly: would the plan differ under either reading? If both readings lead to the same next cycle, a third opinion buys anxiety, not information.

2. 2

   Send images, not the report

   Request the DICOM files from the imaging centre. A reviewing radiologist can re-measure from images. Nobody can re-measure from a scanned PDF or a photograph of a printout.

3. 3

   Send the earlier scans too

   Response assessment is a comparison, not a snapshot. A reader without the baseline and the interim studies is guessing at the very number the two reports disagree about.

4. 4

   Ask for the criteria to be named

   Request that the new report state whether RECIST 1.1, iRECIST or PERCIST was applied, and which lesions were tracked. Two reports naming their criteria are far easier to reconcile.

5. 5

   Take it to a tumour board, not to a fourth centre

   Each new reader introduces the same natural variation that caused the confusion. A multidisciplinary review of the existing studies usually settles the question faster, and without another scan.

Families funding treatment from abroad often receive only the report text on a phone message, which is the hardest possible way to judge a disagreement. Ask for three things in one request: the DICOM images, the prior studies for comparison, and the criteria used. All three usually travel as a single secure link from the imaging centre.

Weigh a repeat scan honestly before paying for one. Repeat imaging costs money, adds radiation exposure and delays the decision by days. Response-assessment PET-CT is coordinated at partner imaging centres; CION's published whole-body PET-CT rates are ₹9,999 for analog and ₹14,950 for digital — indicative, as of August 2026. Repeating a study is worth it when the original was technically inadequate, or when a confirmation scan is due anyway.

Ask for one consultation with both reports open. A 45-minute consultation, with a relative joining by phone, resolves more than three separate opinions collected in three cities. Bring dates: when each scan was done, when treatment started, and which cycle you are on. Timing is often the piece that explains the whole disagreement.

• Pseudoprogression: When the Tumour Looks Bigger but Treatment Is Working — the single most common reason an immunotherapy scan report can read worse than the patient looks.
• When Is the First Scan on Immunotherapy, and Why the Wait? — why scanning too early is one of the reasons two readers end up arguing about millimetres.
• 'Stable Disease' on Immunotherapy: Is That Good News? — the category two reports most often land on either side of, and what it actually means.
• Immunotherapy at CION Cancer Clinics — how consultations, tumour-board review and imaging coordination fit together across the treatment journey.

This page explains general radiology reporting terminology for education only. It does not interpret any individual patient's scan or report, and it is not a substitute for a consultation. Response-assessment PET-CT is coordinated at partner imaging centres, which perform the scan and issue the radiology report. Bring both reports and the images to a consultation for a doctor's assessment.

Families read conflicting scan reports with us every week

A report that seems to say the opposite of how you feel is frightening. Our oncologists take the time to explain, calmly, what the two readings actually differ on.