Multimodality image fusion is a software process whereby images from different sources (CT, MRI, PET, US, Bone Scan) can be overlayed to supplement anatomically enhanced images using varying tissue contrast with functionally enhanced images of physiological/pathological function. The process is based on the ability to present and annotate image profiles in a common data orientation.
Most planning software is currently limited in its ability to provide image fusion. The Eclipse and XiO products will only handle 2 image sets, while Pinnacle will fuse any number of images by only overlay 2 at a time. In order to use an image fusion for clinical means, there must be adequate co-registration of landmarks, and the patient's position must be similar (identical positioning is the optimal, with altered positions being increasingly unhelpful). In the case of PET and bone scans, this can be difficult without indexed couches or the use of markers. US and Bone scans are rarely fused currently, although there are software in development which will permit this.
Not withstanding these caveats, image fusion has the potential to provide more accurate definition of gross tumour and clinical target areas, thereby assisting the oncologist in their role of determining anti-anatomy (the diseased parts) and anatomy (the risk parts) which underpins the planning process. Since planning is best undertaken on the image modality providing electron density measures (the CT), the use of overlayed imaging from MRI can allow the finer delineation of soft tissues, especially neural and somatic structures. This better delineation will facilitate more accurate delineation of at risk organs such as cranial nerves, as the oncologist is better able to apply their detailed anatomical knowledge. It should be noted that while image fusion can allow the GTV and CTV to be more precisely defined, it is of little assistance in generating a PTV which is after all a geometric construct and not drawn by oncologists.
Conventional CT-based planning is something of a misnomer as "conventional" is usually used to describe an outdated simulation method where field borders were placed on plain radiographs at a time when the tools to delineate soft tissue boundaries were limited. All CT-based planning should be dosimetric and based on real soft tissue anatomy. The subsequent 3DCRT and IMRT approaches to planning to cover the defined volumes are based on the identical basis to anatomical voluming at simulation, with the difference residing in the method of planning the resultant covering isodose patterns.
Image fusion has enhanced all forms of planning by improving the anatomical definition and pathological identification of tissues to be treated.
For instance in the case where a submandibular glans is involved with adenoid cystic carcinoma showing perineural infiltration, post-operative radiotherapy can be targeted specifically at the course of the hypoglossal nerve as it passes posteriorly inside the mandible and over the medial pterygoid muscle towards the skull base. The nerve is well seen on an MRI, better than on a CT. The oncologist is then able to deliver a tightly confined dose to a greater length of the nerve without the attendant mucosal toxicity of a large conventional field approach.
In Head & Neck cancer radiotherapy, the PET scan has the ability to identify small but positive lymph nodes that should be treated with higher doses, but would otherwise be missed. It also has the ability to delineate the extent of invasion into deep structures which is particularly useful in organs such as the tongue where infiltration can be subtle, and the mandible where extension through the bony trabeculae can be under-appreciated.
In prostate cancer radiotherapy, MRI is more useful than the CT as the apex of the prostate and presence of surrounding pelvic floor muscle and vascular structures can be better appreciated. This modality also reveals some of the potential problems of image fusion. It is possible to delineate the seminal vesicles (SV) from their origin (unlike on the CT), and so it is possible to exclude the SV from the target area without realising that the lower portion of the SV, which could not be identified until they separated from the prostate, have always been treated, and so without specific evidence cannot just be excluded from the target volume.
Application of multimodality image fusion in radiotherapy treatment planning:
- The integration of imaging data from multiple sources (MRI, PET scans) aids in a more precise definition of tumour volumes due to the additional anatomical or biological information. A better definition of tumour volume allows dose escalation
(ALLOWS FOR DOSE CONFORMITY FIRST, AND THE POSSIBILITY OF ESCALATION SECOND)
* MRI provides superior soft tissue contrast relative to CT and allows visualisation and delineation of certain anatomic structures such as optic nerves and chiasm.
* The fusion of T1, T2 and brainlab sequences in MRI of the brain with the CT dataset aids delineation of cerebral neoplasms. T1 weighted images with gadolinium are used for the delineation of tumour volume in glioblastoma multiforme and T2 weighted flair images are used to delineate the tumour volume and oedema for low-grade gliomas.
* The fusion of the PET-CT scans (ideally in the same RT treatment planning position) incorporates the functional information from PET scans and aids in the delineation of tumours/lymph nodes with FDG-18 uptake. These include tumours such as lung cancers, head and neck cancers, oesophageal cancers and gynaecological malignancies. Functional information from the PET-CT scans may remove
the ambiguity that may
(IS IT POSSIBLE THAT THIS AMBIGUITY DOES NOT EXIST?)
exist on the treatment planning CT between tumour and other conditions such as atelectasis and necrosis
- Reduces inter-observer variability in delineation of tumour volumes and improves standardisation of treatment volumes.
(I DON'T SEE HOW THIS OCCURS. IF MY CT ANATOMY IS CRAPPY, IT IS LIKELY THAT MY MRI ANATOMY WILL ALSO BE CRAPPY. IT IS ALSO UNLIKELY THAT THE CRAPPINESS OF MY CT ANATOMY WILL EXACTLY MATCH THE CRAPPINESS OF YOUR CT ANATOMY ALSO, SO HOW WILL THAT STANDARDISE?)
- The fusion of PET-CT scans takes into account organ motion and allows a tighter margin for planning target volumes
(I THINK THAT YOU NEED TO PHRASE THIS DIFFERENTLY. THE EXECUTION OF MULTIPLE IMAGES HAS THE POSSIBILITY OF REVEALING THE EXTENT OF MOVEMENT SO AS TO INFLUENCE THE CTV-PTV MARGIN. IF A MRI SHOWS THAT THE PROSTATE HAS MOVED POSTERIORLY, THE ONCOLOGIST CAN CHOSE TO USE UNIFORM MARGINS AROUND THE CT-CTV FOR THE PTV (SO LONG AS IT ENCOMPASSES THE MRI-CTV COMPLETELY), OR PRODUCE A COMBINED CT/MRI-CTV WITH A REDUCED CTV-PTV MARGIN BECAUSE SOME OF THE MOVEMENT HAS BEEN REVEALED)
* As the PET-CT image is acquired over several minutes, the image obtained from the PET-CT incorporates organ motion
(BUT THE PERIPHERAL MARGINS WILL FADE INTO BACKGROUND AND SO SUV CHOICE BECOMES CRITICAL)
An example of this is in lung cancer. The PET-CT scan taken over several minutes will incorporate lung motion. Thus, it allows a smaller expansion for organ motion in the planning target volume.
Structure is there but discussion is light on. In particular you are asked for examples where fusion influences CT-based planning. You should have some specific ones at hand that make them think that you actually know what you are doing, wonderful for throwing in some anatomical knowledge (like the CN XII anatomy in the upper reaches where most oncologist's knowledge will be rusty). The use of fusion scans to detect motion is somewhat controversial and a secondary use, so you should spend a little time giving them a sense that you know what you are doing - you need to describe a logic for them that makes sense. You need to use the terms GTV and CTV like you know how to use them effortlessly, and to deliberately fob off PTV construction.
There are lots of places where:
- better identification of the optic apparatus can lead to do dose painting to prevent neuropathy.
- functional image can
- direct high dose boosting of unenlarged LNs
- change intent of therapy from curative to palliative and vice versa
- better anatomical delineation of involved sites can point to a better technique due to proximity of poorly defined critical structures