The past 40 years have witnessed an explosion in medical imaging. It now impacts on all cancer patients at many points on their pathway. Traditionally used to guide diagnosis and staging and determine treatment response, it is now also used to deliver treatments with both palliative and curative intent. This growth in imaging offers many tests—often more than one for the same indication. Preferred tests are influenced by competing factors, including accuracy, cost and availability (Table 1).
Table 1 Overview of imaging tests in oncology.
The risks of radiation exposure are also important. Medical/dental imaging now accounts for 90% of all artificial radiation. Estimates suggest that 1.5– 2.0% of cancers in the USA may be attributable to computed tomography (CT). Although in cancer patients the individual risk may be commuted by limited prognosis, greater consideration should be afforded to children, given their longer life expectancy and increased sensitivity to radiation (2–5 times that of adults).
Staging is the process of determining the location and extent of a malignant disease. Now in its seventh edition (2010), the American Joint Committee on Cancer (AJCC) sets out through an evidence-based review of impacts on survival how clinicians, radiologists and pathologists stage cancers with regard to the primary tumour (T), lymph nodes (N) and distant metastases (M). Imaging tests typically perform best at one or two of these subsections, but not all three; hence the frequent need for multiple staging investigations.
Plain-film radiography in cancer care has largely been usurped by other modalities. The most common cancer indications include the first detection of lung and bone tumours and the investigation of symptoms during treatment or follow up (e.g. neutropenic sepsis and drug-related lung injury. Chest x-ray (CXR) is a relatively insensitive and nonspecific test that has been repeatedly investigated as a screening tool for lung cancer but not shown to be effective (see Figure 2). However, plain radiographs are normally the initial investigation of choice due to their low cost and wide availability.
Figure 2 CXR (a) was requested for shortness of breath. This raised concern over the right hilum, resulting in CT (b), which confirmed the hilum as normal but revealed a CXR occult lung carcinoma (arrow). PET-CT (c) showed stage T1aN0M0 (arrow), so this was treated with curative resection.
Fluoroscopy uses x-rays in real time rather than as a still image. It now uses sophisticated devices deployed by interventional radiologists to guide stent placement, in order to relieve malignant obstructions of the bowel, bile duct and veins, as well as to deliver targeted therapy such as radioactive microspheres to liver tumours. In addition, it is used for the placing of lines or ports, enabling delivery of systemic treatment.
Mammography remains the imaging of choice for breast screening in the UK. It is better at detecting microcalcification (MCC) that are the ultrasound and magnetic resonance imaging (MRI), which are superior in assessing the density of a breast lesion and, in the case of MRI, enhancement characteristics.
Mammography uses conventional x-rays but is moving to digital use rather than x-ray film. In digital mammography, the conventional film is replaced with solid-state detectors that convert x-rays into electrical signals, which then produce computer-generated images of the breast.
The move from analogue to digital has improved the spatial resolution of mammography and has allowed tomograms of the breast (similar to CT), a technology known as tomosynthesis. This latest technique is yet to find regular clinical application but early studies suggest that it allows composite shadowing to be clearly distinguished from a mass or distortion with greater confidence. It does not, however, improve on the detection or characterisation of MCC, where mammography remains superior. While mammography is better for detecting MCC and therefore ductal carcinoma in situ (DCIS), tumour extent is better delineated with MRI, where tissue enhancement is key.
Diagnostic ultrasound utilises probes of frequencies varying between 2 and 18 mHz. These probes act as transmitters and receivers of the returning sound waves in order to generate an image. In general, the higher the frequency, the better the resolution but the poorer the penetration. The converse is also true.
Ultrasound is particularly good for assessing the solid organs, especially the liver, where it plays a significant role in staging disease, for example in patients with breast cancer, or as surveillance in detecting primary liver tumours in patients with chronic liver disease, as well as determining indeterminate low density lesions detected on a staging CT. It is very good at distinguishing solid from fluid-filled lesions and has fine resolution; lesions only a few millimetres in diameter can be outlined clearly. Ultrasound is limited by body habitus, where large patients reduce the penetration of the sound waves and thus image resolution.
Endoscopic ultrasound (EUS), where an ultrasound probe is attached to an endoscope, has proved invaluable in assessing structures in the chest (endobronchial ultrasound, EBUS) and abdomen, which can be obscured by gas; as the lesions are closer to the probe, it also offers a significantly better resolution. It helps with assessing the bowel or airway wall, pancreatic/retroperitoneal lesions and lymph nodes. Another advantage is that lesions can be sampled at the time of examination.
The advent of microbubbles, a contrast agent for the ultrasound, has significantly improved the sensitivity and specificity for the detection and characterisation of focal liver lesions so that they rival those of CT and MRI. In particular, contrast-enhanced ultrasound has a very high sensitivity in the detection of metastases where they appear as defects in the late phase; that is, past the portal phase of enhancement at around 2–3 minutes post-injection. Lesions as small as 3–5 mm can be accurately characterised.
Owing to its wide availability, ultrasound remains the initial imaging modality when it comes to assessing the solid organs in the abdomen. Its easy manouvrebility and good spatial resolution, particularly with superficial and small structures, also make it the modality of choice for the assessment of palpable ‘lumps and bumps’. It is the most utilised modality for image-guided biopsy.
CT is the foundation test for many cancer patients. The majority of cancers are staged with CT and it is the most common imaging test for radiotherapy treatment planning, response assessment, surveillance (when appropriate) and suspected relapse. It provides an excellent overview of disease in an economical (NHS tariff for body CT = £144), readily available and rapid test (a whole-body CT takes just seconds to acquire). It also acts as a ‘gatekeeper’ to more costly, less available complimentary modalities; for example, in oesophageal cancer staging, it provides an overview of disease extent. If CT reveals no distant disease, EUS is used for more accurate primary (T) and lymph node (N) staging and positron emission tomography-computed tomography (PET-CT) as a more sensitive search for metastases (M). Where greater tissue contrast is required to separate tumour from normal structures, MRI outperforms CT; for example, in the staging of rectal, uterine and cervix cancers. An emerging role for CT is cancer screening. The bowel cancer screening programme uses CT colonography when colonoscopy is unacceptable or incomplete. The National Lung Screening Trial (NLST) was stopped prematurely after it revealed that low-dose CT screening of high-risk individuals reduced disease-specific mortality by 20%. The results of a UK trial are awaited and a cost-benefit analysis is required before it is adopted by the NHS.
Magnetic resonance imaging
MRI technology is advancing, with larger field strength magnets providing improved image quality, particularly with 3 Tesla (T) MRI scanners. The majority of clinical scanners at present work at 1.5 T. MRI is absolutely contraindicated in patients with pacemakers and there is still concern regarding the safety of scanning patients with metalwork at higher field strengths.
MRI provides very good soft tissue differentiation when compared to CT or ultrasound but is hampered by relatively long scanning times and therefore movement artefacts. Scan times are getting progressively faster, with stronger gradients and advancing technology.
Altering sequences and tissue weightings can help better evaluate different types of tissue and advancing techniques such as diffusion-weighted imaging (DWI) and magnetic resonance spectroscopy (MRS) can provide an insight into the metabolic activity of lesions. The restriction of water mobility within a cellular environment in DWI is thought to represent an early harbinger of biological abnormality. These additional DWI sequences may be useful in assessing the aggressiveness of lesions and in tumour response. MRS, on the other hand, gives information on chemical cellular content where graphs are obtained. The graph profiles of malignant and benign tissues will differ due to their contrasting tissue compositions. This currently remains a research technique.
As with the other cross-sectional imaging techniques, contrast enhancement with intravenous gadolinium-based agents can help improve tissue differentiation and assess the vascularity of lesions. There are also contrast agents with a reticulendothelial uptake phase and some which are biliary excreted; therefore, the examination can be tailored to suit the organ or clinical scenario of choice.
In cancer imaging, MRI is not yet typically used for whole-body staging; this is partly due to cost, but it also has limited accuracy in lung lesions when compared with CT. Typically, MRI is used to assess locoregional spread and stage of the disease. For example, in colorectal, prostate or gynaecological cancers, MRI provides highly accurate tumour extent and lymph node staging.
MRI has proved invaluable in imaging the central nervous system (CNS) and spine for malignancy and particularly cord compression, where it remains the modality of choice over the other cross-sectional techniques.
Overall, this field continues to advance rapidly, and whole-body MRI is now possible. In some instances, it has been used as a screening tool or in the detection of metastases and disease response, but its clinical value and efficacy in this respect remain a subject of study.
Nuclear medicine studies function at a molecular level rather than through changes in anatomy. There are a plethora of tests, which rely on increasingly specific radiopharmaceuticals to image or treat cancer patients. Gamma and PET cameras have relatively poor image resolution, so are increasingly combined with CT and more recently MRI scanners to create fused functional and anatomical data. Despite the unique potential for individually tailored imaging agents, the majority of nuclear medicine tests performed in cancer patients are technetium 99m bone scans for the detection of metastases and 18F-FDG PET-CT, which observes nonspecific glucose uptake by highly metabolic cancer cells.
Bone scanning relies on osteoblastic activity, so is most accurate when searching for sclerotic bone metastases, from the prostate and breast, for example (see Figure 2). False positives may occur when scanning patients in the recovery phase of treatments, whereby healing tissue may be mistaken for active disease. MRI and PET-CT are superior to bone scanning where lesions are lytic, such as in renal cell cancer. MRI is used in preference to the other modalities where spinal cord/nerve compression is suspected.
Figure 2 Prostate cancer with disseminated bone metastases on bone scan (a). Later presentation with sacral nerve root signs. MRI (b,c) revealed both the extensive bone marrow infiltration and the sacral foraminal narrowing (arrow).
PET-CT infrequently improves T-staging over CT or MRI and, while better for N-staging (CT N-staging in lung cancer: sensitivity 51%, specificity 86%), remains insufficiently accurate to determine management alone (10% false-negative rate in lung cancer), ensuring a role for mediastinal sampling in presurgical staging. Where PET-CT outperforms other tests is in metastatic detection. Among patients staged by CT, PET-CT prevents futile thoractomy in 20% (Figure 3).
Figure 3 PET-CT (a,b) staging for lung cancer. The large CT-evident mediastinal lymph node (N2) was intensely FDG avid. PET also revealed an avid supraclavicular fossa lymph node (N3) (arrow) not apparent on CT (c), even in hindsight, as well as an intramuscular metastasis (arrow) also occult on CT. Stage migration between tests was T2aN2M0 to T2aN3M1b.
To gain some understanding of how these tests interleave in a patient’s cancer journey, two typical pathways are outlined in Figure 4.
Figure 4 Algorithm of typical patient pathways for lung and colorectal cancer, illustrating where imaging may be used and the preferred imaging modality.