Monday, February 24, 2014

Pixels aren't cells: the benefits and limitations of imaging in the staging of oncology patients

Today's post features Dixie, who doesn't have any medical conditions but enjoys sitting up on the couch and slowly destroying paper bags.
Last week we attended a great seminar by Animal Referral Hospital imaging specialist Dr Karon Hoffman on imaging for staging of oncology patients. Dr Hoffman is a Diplomate of the European College of Diagnostic Imaging, as well as a recipient of the CVE’sHungerford Award. She has both masters and PhD qualifications and is yet another person on our long list of people we want to be when we grow up.

It is always fascinating hearing how specialists of this calibre organise their thoughts. Staging, of course, is the determination of the anatomical spread of cancer and imaging is one part of it (as is physical examination, but also occasionally surgery). It is different from grading, which is performed by pathologists.

So once you have detected a tumour, why perform staging? In the most basic sense it is to determine tumour size, the extent of invasion and margins, involvement of regional lymph nodes and identification of metastases.

Much of this is achieved through imaging, which is also important for surgical planning. Fortunately some tissues (fascial planes, tendons, cartilage and bone) are more resistant to invasion by neoplastic cells than others but even benign tumours recur with incomplete resection so we want to achieve clean margins the first time. Knowing where to cut is vital!

Dr Hoffman discussed the pros and cons of different imaging modalities commonly used in the staging of veterinary oncology patients.

Just like a work of art...Dixie strikes yet another thoughtful pose.
Advantages: provides global information, is relatively inexpensive, especially useful for imaging the appendicular skeleton.

Disadvantages: superimposition of structures, may not be read in detail (Dr Hoffman said that our eyes tend to really see an area of the radiograph about the size of a 20 cent piece…so if we don’t cast our eyes over the whole film we’re missing a lot), increasingly replaced with CT and MRI (especially for the head and axial skeleton).

Interestingly, for a metastases hunt, Dr Hoffman suggested that sometimes four views of the thorax are helpful: both laterals plus ventrodorsal and dorsoventral projections. While most publications recommend three views, Dr Hoffman has worked in some centres where four thoracic views is the standard.

There are two potential exceptions to the three (or four) radiograph requirement: lymphoma (usually thoracic lymph nodes, pleural effusion and pulmonary infiltration can be appreciated on two lateral views) and mast cell tumours which rarely metastasize to the lungs.

Of course, evaluating the patient for concurrent disease is important for staging and prognostication, but according to Dr Hoffman abdominal ultrasound would be a better investment of resources for dogs with MCTs than thoracic ultrasound.


Advantages: excellent for evaluation of soft tissue and fluids, provides guidance for fine needle aspiration and biopsy, useful for monitoring response to treatment (e.g. tumour size), often won’t need a general anaesthetic. Ultrasound is excellent when the operator is methodical, thorough and experienced.

Weaknesses: operator dependent (you will miss the lesion if you don’t point the transducer at it), provides local (not global) information, 3-dimensional interpretation is in the operator’s mind.

Sonography is less helpful in the case of osteosarcoma (OSA) as these generally metastasize to the lungs, however extraskeletal OSAs occur and can be detected on abdominal ultrasound.

Dr Hoffman spent a lot of time discussing what can and cannot be inferred from imaging. As veterinarians we are frequently asked to take a best guess at what a lump is likely to be – but owners will often use this information to make life-or-death decisions.

Nodules on the liver or spleen are common but Dr Hoffman said that many animals have died because these are found on ultrasound. Such nodules may represent metastases, but they may also be benign lesions.

Without a biopsy it is difficult to make an assessment. One might suspect that these are neoplastic but further evidence should be sought.

She cited Aristotle: “It is the mark of an educated mind to be able to entertain a thought without accepting it”.

In other words, consider neoplasia as a differential, even the most likely differential if it fits the clinical picture, but don’t seal the deal without further evidence.

Splenic masses are a case in point. We tend to fear that every splenic mass represents a haemangiosarcoma, but this accounts for only 1/3 of splenic masses. Another 1/3 are non-neoplastic, and the remaining 1/3 are tumours of some other kind, 50% of which are malignant. Which means that 50 per cent of splenic masses are curable!

Computed tomography

Advantages: avoids the problem of superimposition by taking imaging “slices” of the subject, and utilises various algorithms to optimise contrast, provides 3-D image for surgical planning and may help avoid abortive surgical attempts, very sensitive in detecting lung lesions (CT can detect lesions of 1mm in diameter, radiology can only detect lesions from 4-7mm in diameter - though the prognostic significance remains unknown since most literature on prognosis involves data based on radiographic assessment for metastases). CT is superior when compared to MRI in detecting the size of length of OSA

Disadvantages: costly, available in fewer centres, requires general anaesthesia.


Advantages: provides cross sectional imaging, excellent to illustrate tumour margins and metastases, sensitive for detecting inflammation.

Disadvantages: slower than CT, costly, available in fewer centres, requires general anaesthesia.


Advantages: excellent screening for lesion detection (especially skeletal metastases and thyroid cancer), permits functional studies.

Disadvantages: low resolution, available in fewer centres.

Her take home advice was this: “People want to know what a tumour is based on the gray scale on the screen. The pixels aren’t cells and while we need a gross structural idea of where the tumour has gone we also need to know the cell line and the characteristics of that cell line.”