A Case Of Nasopharyngeal IMRT

This person (non-Asian origin) produced with transient cranial nerve palsies and neck nodes. A PET scan showed hot areas in the nasopharynx, clivus and neck nodes bilaterally. Staging was T4N2M0, non-keratinising SCC.

The planning CT was volumed for a GTV (CT-visible disease) according to the IMRT protocol of the department. The GTV was expanded by a 3mm margin into normal tissues to form a CTV. The PET scan was fused with the planning CT. The CTV was expanded to encompass areas that were hot on the PET scan but not obviously abnormal on the CT, and in these areas expanded by 3mm so that the entire area of CT and PET GTVs were expanded by 3mm. The clivus volume is not symmetric because the clivus was not entirely involved, just crossing the midline at maximum extent with the contralateral clivus being cold on differing PET SUV settings. Because the 'correct' SUV setting is unknown, the SUV threshold was varied to give an impression of the extent of abnormality in relation to the background count.

The protocol describes the formation of critical structures which fall into only one pattern for this case, because the disease was encroaching the CNS. Irradiation of the CNS is unavoidable, so the emphasis then moves to reducing the amount of CNS tissue irradiated. Normally the CNS structures are expanded by the designated CTV>PTV margins (internal movement and setup margin) as they are critical structures (A structure which has a notional maximum dose is called a "critical structure", while those that are subject to attempts to reduce dose are called "organs at risk". This distinction is arbitrary but has ramifications for planning).

In this case all organs at risk move in a manner the same as the CTVs, so all OARs are expanded to produce PRVs. In addition, in a perfectly conformal plan, the 95% isodose will huge the PTV outline like Spandex, and there will be 4-6mm penumbra where the dose falls from 95% to ~60%. This penumbra is inevitable and unavoidable, and you can try anything you like but the physical realities are that it will be there when you are finished! Thee is a way to take this into account.

The CTVs are expanded to produce the PTV using the expected movement. With a shell alone this would be 5mm, but our department uses an image guidance protocol that assesses the isocentre each day. In fact the isocentre is placed anatomically by the RTs to give an easy to assess isocentre (on C4). For this case the isocentre is to be verified each day, i.e., a margin of 0mm. Some unpublished Australian data suggests that this can be achieved to a 1mm accuracy, so the CTV>PTV expansion is 1mm. The reason for the altered protocol in this case relates to fact that each additional mm of movement will increase the dose to the CNS.

The PTVs are produced with a dose label. PTV60 includes all the high risk CTVs (CTVp & CTVn) and low risk CTV (CTVn0), while the PTV70 includes all the high risk CTVs (CTVp & CTVn).

So far, nothing really radical. Now the different stuff! The PTV60 is expanded by 7mm to produce a PTV_PRV. The OAR is then expanded by the expected movement and excluded at the PTV_PRV. You can see this on the slice containing the parotid outlines. Although the conventional wisdom is to aim for a median dose of 26Gy to the parotids, I find this thinking deficient - what if the median dose could be less than this? Shouldn't the parotid dose be minimised?

So the parotid_PRV is limited to 18Gy even less of the plan permits. In this case the median parotid doses are 25Gy and 18Gy! Better than most people aim for! This pattern of OAR_PRV generation is undertaken with all the OARs (CNS, parotids & submandibular glands, cochleas, mandible and mucosa) and absolute doses assigned to each PRV.

Then the planner goes to work translating those criteria into priorities for the IMRT optimisation engine to use to generate an acceptable plan. Now I should warn you that the planner did not like this plan! We had to fiddle a little to get a reasonable compromise. The mucosa_PRV dose was relaxed al little, the mandible_PRV was decreased by about 1cm to allow more leeway to preserve parotid_PRV dose levels.


This image shows the dose high up in the volume at the clivus. The volumes are in thick lines and the isodoses are thin lines. RED is the GTV, ORANGE is the CTV and thick PURPLE is the PTV70 (the thick PURPLE PTV70 is hidden by the thick BLUE PTV60 - they are exactly the same size). The 95% of 70Gy isodose (6650cGy) is GREEN and the 95% of 60Gy isodose (5700 cGy) is PURPLE. The other lines are the LEMON-GREEN (6000cGy) and YELLOW (5000cGy).


In this image you can now see the thick PURPLE and BLUE PTVs, as well as the shaded PAROTID_PRVs (purple). You can see how the isodoses are pushed inwards by these structures. You can also see the node of Rouvier on the body of C2 which was hot on PET scan. This is contained within the PTV70 volume and is receiving >95%.


This image is further down the neck and demonstrates the effect of generating a MUCOSA_PRV which, in this case, is limited to 50Gy (but usually 40Gy). You will also notice that there are PTV70 lines included within the PTV60. All visible nodes in the neck are volumed and included in the PTV70.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License