Clinical dosevolume histogram analysis
for pneumonitis after 3D treatment for non-small cell lung cancer (NSCLC)
Graham MV, Purdy JA, Emami B, Harms W, Bosch W, Lockett MA, Perez CA
IJROBP 199;45:323-329
Between January 1991 and October 1995, 99 patients were
treated definitively for inoperable NSCLC. Patients were selected for good performance
status (96%) and absence of weight loss (82%). All patients had full 3D treatment planning
(including total lung dosevolume histograms [DVHs]) prior to treatment delivery. The
total lung DVH parameters were compared with the incidence and grade of pneumonitis after
treatment.
Results: Univariate analysis revealed the percent of the total
lung volume exceeding 20 Gy (V20), the effective volume (Veff) and the total
lung volume mean dose, and location of the tumor primary (upper versus lower lobes) to be
statistically significant relative to the development of ? Grade 2 pneumonitis.
Multivariate analysis revealed the V20 to be the single independent predictor of
pneumonitis.
Until further investigations of this nature can be performed, we recommend that the
total lung volume DVH be assessed when evaluating the goodness of a 3D
radiation plan in the treatment of NSCLC patients. In our clinic
when the total lung V20 is <25%, we are comfortable with tumor dose escalation and the
very low risk of pneumonitis. These plans are considered
acceptable.
If a plan has a total lung V20 of >25% to 37%, alternative
plans are done with an attempt at reducing the V20. This may be achieved by
different beam arrangements, noncoplanar beams, less or no elective nodal irradiation, or
smaller margins around the target volumes. This last technique is done only as a last
resort and should be carried out with great caution as it may decrease the dose delivered
to the tumor. There is much uncertainty in dose calculation accuracy at the PTV periphery
greater due to the high dose gradient regions at the margin of tumor and low-density lung
tissue.
If a treatment plan gives a V20 of >3540%, we do not use
that plan for treatment. All fatal pneumonitis occurred in patients with a V20
>35%. Similarly, all high-grade pneumonitis occurred in patients with a V20 of >32%.
The risk of pneumonitis, in our estimation, is too great. Options for treatment then
include: (1) changing the plan, as outlined above, (2) administering neoadjuvant
chemotherapy in an attempt to reduce the volume of the tumor and treat the
postchemotherapy tumor volume, and (3) treating the patient palliatively with lower doses.

|