Analysis of Visual Loss Due to Radiation-Induced Optic Neuropathy After Particle Therapy for Head and Neck and Skull Base Tumors Adjacent to Optic Nerves.
Reviewer: Christine Hill, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: May 23, 2008
Presenter: Yusuke Demizu Presenter's Affiliation: Department of Radiology, Hyogo Ion Beam Medical Center, Tatsuno, Hyogo, Japan Type of Session: Scientific
Use of radiotherapy for treatment of tumors of the head and neck is made complicated by the proximity of vital structures, including the optic chiasm and optic nerves. Clearly, visual loss (VL) can have significant impact on quality of life (QOL) after cancer treatment
Because of unique physical properties that can allow improved dose distribution with particle beam therapy, damage to these structures is expected to be lessened with particle beam treatment versus photon radiotherapy; however, clinical data investigating this remains relatively scarce.
Additionally, we have only limited knowledge and understanding of the tolerance dose of the optic nerve with particle beam irradiation.
In 1991, Emami et al described a 5% risk of complications with a total optic nerve dose of 50 Gray (Gy) during photon radiation.
Urie et al subsequently found a 5% risk of complications with a total dose of 70 Gy to the optic nerve using proton beam radiotherapy.
In abstract form, Hasegawa et al have presented complication data, describing increased risk of blindness associated with dose of 60 Gray Equivalents (GyE) or greater to 20% of the optic nerve in the setting of carbon ion radiotherapy.
This retrospective, single-institution study was undertaken in order to evaluate the incidence of visual loss after particle therapy for head and neck and skull base cancers, and to evaluate risk factors for VL in this setting.
Materials and Methods
During the period from August, 2001 to August, 2006, 104 patients underwent particle-based radiotherapy for tumors adjacent to the optic nerve at Hyogo Ion Beam Medical Center.
Proton beam irradiation was delivered to a total dose of 65 GyE in 26 fractions.
Carbon ion radiotherapy was delivered to a total dose of 57.6 GyE in 16 fractions.
Of these patients, 112 optic nerves of 72 patients underwent at least 50% of the prescribed radiation dose, and were followed for at least 12 months.
92 optic nerves had received proton beam irradiation, and 20 had received carbon ion irradiation.
Median age was 61 years, with an age range of 15 – 85 years.
Nasal cavity and paranasal sinus were the most common primary disease sites.
75% of patients/ optic nerves analyzed had received no prior treatment.
Three patients had concurrent diagnosis of diabetes mellitus, and 11 had diagnosis of hypertension.
Visual loss was defined as the inability to count fingers or more severe.
Maximum optic nerve dose was obtained from dose-volume histograms, and the biologically effective doses at α/β = 3 GyE (BED3 with units of GyE3) were compared.
Incidence of VL was determined by the Kaplan-Meier method, and log-rank test and Cox proportional hazards model were employed for univariate and multivariate analyses regarding risk factors for VL.
Median follow-up was 25 months for all optic nerves analyzed. Median follow-up was 28 months in the carbon ion group, and 24 months in the proton group.
Median optic nerve dose maxium was 55.9 GyE3 inthe group treated with proton therapy, and 54 GyE3 in the group treated with carbon iontherapy.
Of the 112 optic nerves analyzed, VL developed in 9 (8%).
Visual loss developed in 10% (n = 2) of optic nerves treated with carbon ion therapy, and 8% (n = 7) of patients treated with proton therapy (p = 0.432).
Median time of onset to VL was 33 months, and ranged from 18 months to 72 months. All cases of VL in patients treated with proton therapy developed less than 14 months after completion of radiotherapy. All cases in patients treated with carbon ion therapy developed greater than 14 months after completion of radiotherapy.
Of patients developing VL, one developed recurrent disease, and another developed bilateral VL.
Recurrent cancer developed in a 66 year old male treated with carbon ion radiotherapy. Maximum dose to the right optic nerve was 58 GyE3, and to the left optic nerve was 57.5 GyE3. VL developed in the right optic nerve immediately prior to diagnosis of tumor recurrence.
Bilateral VL developed in a 78 year old diabetic woman treated with proton radiotherapy. Maximum dose to the right optic nerve was 40.1 GyE3, and to the left optic nerve was 67.7 GyE3. Bilateral, simultaneous VL subsequently developed
In univariate analysis, age greater than or equal to 65 years, presence of diabetes mellitus, and maximum optic nerve dose greater than 110 GyE3 were all significant factors for development of VL.
The authors note that 110 GyE3 is approximately equal to 52 GyE delivered in 16 fractions, or 61 GyE delivered in 26 fractions.
In multivariate analysis, age greater than or equal to 65 years, female gender, and concurrent diabetes mellitus were significant factors in the development of VL.
The authors conclude that increased age, female gender, and concurrent diabetes mellitus may contribute to risk of VL during particle therapy for tumors adjacent to the optic nerves.
They note that maximum optic nerve doses above 110 GyE3 may contribute to radiation associated optic neuropathy. Although they acknowledge that this factor did not contribute significantly to risk in multivariate analysis, they point out the proximity of this maximum dose to that previously identified by Urie and colleagues, as well as Hasegawa and colleagues at this symposium, may increase its validity.
They note that these findings may contribute significantly to treatment planning optimization in patients with tumors near or adjacent to optic nerves undergoing particle therapy, and that particular caution should be used in treatment of elderly patients, those of female gender, and those with concurrent diabetes mellitus.
This retrospective, single institution study provides valuable information with regards to risk factors for VL after radiotherapy for head and neck and skull base tumors.
The authors did identify a dose on univariate analysis that could possibly represent a maximum dose threshold for visual preservation; as the authors note, this data taken in conjunction with the findings of other groups is more robust than it would be alone. In multivariate analysis, increased maximum optic nerve dose did not appear to impact risk of visual loss; however, in the absence of other data, the authors’ findings of 110 GyE3 could represent a possible goal for optimization.
The authors’ observation that VL following proton radiotherapy appeared to occur more rapidly than VL following carbon ion therapy potentially has important clinical implications; however, the very small sample size of patients experiencing visual loss after carbon ion radiotherapy (n = 2) increases the possibility that this temporal relationship is based on chance alone.
As accessibility increases and more patients in turn undergo particle therapy for head and neck and skull base tumors, data in larger cohorts of patients with longer follow-up will certainly be of interest. In the absence of further data, however, the maximum safe optic nerve dose identified on univariate analysis in this study is likely a reasonable constraint to impose during treatment planning.
As the authors point out, extra caution is warranted during treatment of elderly and female patients, as well as those with concurrent diabetes mellitus, based on the findings of this study. It is also important to note that the defined testing for visual loss in this study of counting fingers would not pick up smaller visual disturbances that could significantly impair quality of life. Lesser degrees of vision loss then evaluated in this study could effect the ability to read, drive, and perform other daily tasks. Future studies should have more formal prospective opthalmologic evaluations.