Analysis of respiratory-induced motion trajectories of individual liver segments in patients with hepatocellular carcinoma.

PURPOSE: To analyze the respiratory-induced motion trajectories of each liver segment for hepatocellular carcinoma (HCC) to derive a more accurate internal margin and optimize treatment protocol selection. MATERIALS AND METHODS: Ten-phase-gated four-dimensional computed tomography (4DCT) scans of 14 patients with HCC were analyzed. For each patient, eight representative regions of interest (ROI) were delineated on each liver segment in all 10 phases. The coordinates of the center of gravity of each ROI were obtained for each phase, and then the respiratory motion in the left-right (LR), anteroposterior (AP), and craniocaudal (CC) directions was analyzed. Two sets of motion in each direction were also compared in terms of only two extreme phases and all 10 phases. RESULTS: Motion of less than 5 mm was detected in 12 (86%) and 10 patients (71%) in the LR and AP directions, respectively, while none in the CC direction. Motion was largest in the CC direction with a maximal value of 19.5 mm, with significant differences between liver segment 7 (S7) and other segments: S1 (p < 0.036), S2 (p < 0.041), S3 (p < 0.016), S4 (p < 0.041), and S5 (p < 0.027). Of the 112 segments, hysteresis >1 mm was observed in 4 (4%), 2 (2%), and 15 (13%) in the LR, AP, and CC directions, respectively, with a maximal value of 5.0 mm in the CC direction. CONCLUSION: A significant amount of respiratory motion was detected in the CC direction, especially in S7, and S8. Despite the small effect of hysteresis, it can be observed specifically in the right lobe. Therefore, caution is required when using 4DCT to determine IM using only end-inspiration and end-expiration. Understanding the respiratory motion in individual liver segments can be helpful when selecting an appropriate treatment protocol.

A novel, end-to-end framework for avoiding collisions between the patient's body and gantry in proton therapy.

BACKGROUND: Administration of external radiation therapy via proton therapy systems carries a risk of occasional collisions between the patient's body and gantry, which is increased by the snout placed near the patient for better dose distribution. Although treatment planning software (TPS) can simulate controlled collisions, the computed tomography (CT) data used for treatment planning are insufficient given that collisions can occur outside the CT imaging region. Thus, imaging the three-dimensional (3D) surface outside the CT range and combining the data with those obtained by CT are essential for avoiding collisions. PURPOSE: To construct a prototype for 3D surface imaging and an end-to-end framework for preventing collisions between the patient's body and the gantry. METHODS: We obtained 3D surface data using a light sectioning method (LSM). By installing only cameras in front of the CT, we achieved LSM using the CT couch motion and preinstalled patient-positioning lasers. The camera image contained both sagittal and coronal lines, which are unnecessary for LSM and were removed by deep learning. We combined LSM 3D surface data and original CT data to create synthetic Digital Imaging and Communications in Medicine (DICOM) data. Subsequently, we compared the TPS snout auto-optimization using the original CT data with the synthetic DICOM data. RESULTS: The mean positional error for LSM of the arms and head was 0.7 ± 0.8  and 0.8 ± 0.8 mm for axial and sagittal imaging, respectively. The TPS snout auto-optimization indicated that the original CT data would cause collisions; however, the synthetic DICOM data prevented these collisions. CONCLUSIONS: The prototype system's acquisition accuracy for 3D surface data was approximately 1 mm, which was sufficient for the collision simulation. The use of a TPS with collision avoidance can help optimize the snout position using synthetic DICOM data. Our proposed method requires no external software for collision simulation and can be integrated into the clinical workflow to improve treatment planning efficiency.

Characterization and clinical utility of different collimator shapes in accelerator-based BNCT systems for head and neck cancer.

NeuCure® is the only accelerator-based boron neutron capture therapy (BNCT) system in the world with pharmaceutical approval. Until now, only flat collimators (FCs) on the patient side have been installed. However, in some cases of head and neck cancer patients, positioning the patient close enough to the collimator when using FCs was difficult. Thus, there are concerns about the prolongation of the irradiation time and overdose to normal tissues. To address these issues, a collimator with a convex-extended section on the patient side (extended collimators [ECs]) was developed, and its pharmaceutical approval was obtained in February 2022. This study evaluated the physical characterization and usefulness of each collimator using a simple geometry water phantom model and human model. In the water phantom model, the thermal neutron fluxes at 2 cm depth on the central axis were 5.13 × 108, 6.79 × 108, 1.02 × 109, and 1.17 × 109n/cm2/s for FC(120), FC(150), EC50(120), and EC100(120), respectively, when the distance from the irradiation aperture was kept constant at 18 cm. With ECs, the relative off-axis thermal neutron flux decreased steeply. In the hypopharyngeal cancer human model, the tumor dose changes were within <2%, but the maximum oral mucosa doses were 7.79, 8.51, 6.76, and 4.57 Gy-Eq, respectively. The irradiation times were 54.3, 41.3, 29.2, and 24.8 min, respectively. In cases where positioning the patient close to the collimator is difficult, the use of ECs may reduce the dose to normal tissues and shorten the irradiation time.

Dosimetric effect of six degrees of freedom couch top with rotational setup error corrections in proton therapy.

PURPOSE: To investigate the dosimetric effect of six degrees of freedom (6DoF) couch top with rotational corrections in proton therapy (PT). METHODS: The water equivalent thickness (WET) was measured using a proton beam with a 6DoF couch top and patient immobilization base plate (PIBP) placed in front of a motorized water phantom. The accuracy verification was performed with the beam axis set perpendicular to the 6DoF couch top and tilted in 10° steps from 10° to 30°. Up to 3° rotational correction may be added during the actual treatment to correct the rotational setup error on our system. The measured and calculated values using the treatment planning system were compared. Additionally, the effect of the 3° difference was evaluated using actual measurements concerning each angle on the proton beam range. RESULTS: The WET of the 6DoF couch top and PIBP were 8.5 ± 0.1 mm and 6.8 ± 0.1 mm, respectively. The calculation and the actual measurement at each angle agreed within 0.2 mm at the maximum. A maximum difference of approximately 0.6 mm was confirmed when tilted at 3° following 30° with the 6DoF couch top plus PIBP. CONCLUSIONS: The dosimetric effect of the 6DoF couch top with rotational corrections in PT differs depending on the incidence angle on the couch top, and it increased with the increased oblique angle of incidence. However, the effect on the range was as small as 0.6 mm at the maximum. The amount of rotational correction, the angle of incidence of the beam, and the effect of rotational corrections on the proton beam range may differ depending on the structure of the couch top. Therefore, sufficient prior confirmation, and subsequent periodical quality assurance management are important.

Dosimetric effects of the ipsilateral shoulder position variations in the sitting-positioned boron neutron capture therapy for lower neck tumor.

We aimed to evaluate dosimetric effects of ipsilateral shoulder position variations (ISPVs) in sitting-positioned boron neutron capture therapy (BNCT) for lower neck tumor. The ISPVs were simulated using deformed shoulder images that can simulate arbitrary shape. The dose-volume parameters for the tumor in the rotated shoulder plans considerably varied compared with that for the mucosa. Even in a small number of cases, these differences were clearly observed among patients. The ISPVs in lower neck BNCT have great dosimetric effects.

Determining a methodology of dosimetric quality assurance for commercially available accelerator-based boron neutron capture therapy system.

The irradiation field of boron neutron capture therapy (BNCT) consists of multiple dose components including thermal, epithermal and fast neutron, and gamma. The objective of this work was to establish a methodology of dosimetric quality assurance (QA), using the most standard and reliable measurement methods, and to determine tolerance level for each QA measurement for a commercially available accelerator-based BNCT system. In order to establish a system of dosimetric QA suitable for BNCT, the following steps were taken. First, standard measurement points based on tissue-administered doses in BNCT for brain tumors were defined, and clinical tolerances of dosimetric QA measurements were derived from the contribution to total tissue relative biological effectiveness factor-weighted dose for each dose component. Next, a QA program was proposed based on TG-142 and TG-198, and confirmed that it could be assessed whether constancy of each dose component was assured within the limits of tolerances or not by measurements of the proposed QA program. Finally, the validity of the BNCT QA program as an evaluation system was confirmed in a demonstration experiment for long-term measurement over 1 year. These results offer an easy, reliable QA method that is clinically applicable with dosimetric validity for the mixed irradiation field of accelerator-based BNCT.

Dosimetric effect of set-up error in accelerator-based boron neutron capture therapy for head and neck cancer.

The dosimetric effect of set-up error in boron neutron capture therapy (BNCT) for head and neck cancer remains unclear. In this study, we analyzed the tendency of dose error by treatment location when simulating the set-up error of patients. We also determined the tolerance level of the set-up error in BNCT for head and neck cancer. As a method, the distal direction was shifted with an interval of 2.5 mm, from 0.0 mm to +20.0 mm and compared with the dose at the reference position. Similarly, the horizontal direction and vertical direction were shifted, with an interval of 5.0 mm, from -20.0 mm to +20.0 mm. In addition, cases with 3.0 mm and 5.0 mm simultaneous shifts in all directions were analyzed as the worst-case scenario. The dose metrics of the minimum dose of the tumor and the maximum dose of the mucosa were evaluated. From unidirectional set-up error analysis, in most cases, the set-up errors with dose errors within ±5% were Δdistal < +2.5 mm, Δhorizontal < ±5.0 mm and Δvertical < ±5.0 mm. In the simulation of 3.0 mm shifts in all directions, the errors in the minimum tumor dose and maximum mucosal dose were -3.6% ±1.4% (range, -5.4% to -0.6%) and 2% ±1.4% (range, 0.4% to 4.5%), respectively. From these results, if the set-up error was within ±3.0 mm in each direction, the dose errors of the tumor and mucosa could be suppressed within approximately ±5%, which is suggested as a tolerance level.

Dose-volume comparison of intensity modulated proton therapy and volumetric modulated arc therapy for cervical esophageal cancer.

Proton therapy for cervical esophageal cancer has many issues to be considered, such as the physiological curvature of the spine and the large range change from the neck to the trunk. We clarified the dosimetric characteristics of intensity modulated proton therapy (IMPT) for cervical esophageal cancer by comparing with volumetric modulated arc therapy (VMAT). Ten patients with cervical esophageal cancer were retrospectively planned for VMAT, 2-field IMPT (2F-IMPT), and 3-field IMPT (3F-IMPT). All plans were optimized to reach clinically acceptable levels. For planning target volume (PTV) coverage, 95% of the PTV should be covered by 95% of the prescription dose, unless the spinal cord limit is violated. The organs at risk included the lung, spinal cord, larynx, skin, and whole body. The prescription dose was 60 Gy relative biological effectiveness (RBE) in 30 fractions to the PTV. We compared the results according to dose-volume metrics. Significant dose reductions were achieved at lung doses, especially at low dose volumes of 20 Gy RBE or less in IMPT plans compared with VMAT plans (p < 0.05). Although the spinal cord PRV was below the tolerance level, the results were also significantly higher in VMAT plans than in IMPT plans (p < 0.001). Spinal cord PRV Dmean was significantly higher in 3F-IMPT than in 2F-IMPT (p < 0.001). In addition, it was confirmed that the integral whole body dose can be dramatically reduced in IMPT plans compared with VMAT plans. Both of 2F-IMPT and 3F-IMPT could effectively reduce spinal cord dose, as well as low integral whole body dose to a certain extent, while maintaining similar target coverage compared to VMAT. IMPT could be a promising treatment technique for patients with cervical esophageal cancer.

Dosimetric properties of a newly developed thermoluminescent sheet-type dosimeter for clinical proton beams.

PURPOSE: This study aimed to evaluate the dosimetric properties of a newly developed thermoluminescent sheet-type dosimeter (TLD-sheet) for clinical proton beams. MATERIALS AND METHODS: The TLD-sheet is composed mainly of manganese doped lithium triborate, with a physical size and thickness of 150 mm × 150 mm and 0.15 mm respectively. It is flexible and can be cut freely for usage. The TLD-sheet has an effective atomic number of 7.3 and tissue-equivalent properties. We tested the reproducibility, fading effect, dose linearity, homogeneity, energy dependence, and water equivalent thickness (WET) of the TLD-sheet for clinical proton beams. We conducted tests with both unmodulated and modulated proton beams at energies of 150 and 210 MeV. RESULTS: The measurement reproducibility was within 4%, which included the inhomogeneity of the TLD-sheet. The fading rates were approximately 20% and 30% after 2 and 7 days respectively. The TLD-sheet showed notable energy dependence in the Bragg peak and distal end of the spread-out Bragg peak regions. However, the dose-response characteristics of the TLD-sheet remained linear up to a physical dose of 10 Gy in this study. This linearity was highly superior to those of commonly used radiochromic film. The thin WET of the TLD-sheet had little effect on the range. CONCLUSION: Although notable energy dependences were observed in Bragg peak region, the response characteristics examined in this study, such as reproducibility, fading effects, dose linearity, dose homogeneity and WET, showed that the TLD-sheet can be a useful and effective dosimetry tool. With its flexible and reusable characteristics, it may also be an excellent in vivo skin dosimetry tool for proton therapy.

End-to-end test to evaluate the comprehensive geometric accuracy of a proton rotating gantry using a cone-shaped scintillator screen detector.

In this study, we aim to evaluate the comprehensive geometric accuracy of proton rotating gantries by performing an end-to-end test using a cone-shaped scintillator screen detector, known as XRV-124. The XRV-124 comprises a cone-shaped sheet-like scintillator and charge-coupled device camera that detects the scintillation light. First, the results of the Winston-Lutz and end-to-end XRV-124 tests performed on a conventional linear accelerator were compared to confirm the reliability of the XRV-124, and the snout position dependency of the geometric accuracy was evaluated for the proton rotating gantry as a pre-verification process. Thereafter, an end-to-end test including computed tomography imaging and irradiation in 30° steps from 0° to 330° for two proton rotating gantries, which have the same specifications, was performed. The results of the pre-verification indicated that sufficient accuracy was obtained for the end-to-end test of the proton rotating gantry. The end-to-end test results showed a peak-to-peak deviation of up to 2 mm for some of the coordinate axes. The two gantries exhibited almost similar results in terms of the absolute quantity; however, a few trends were different. Thus, the beam axis deviations were confirmed to be within the safety margin, as expected in clinical practice. Based on the results of this study, the XRV-124 can be used as a comprehensive end-to-end constancy test tool, as it enables a comparative verification of multiple rotating gantries and geometric accuracy verification of different treatment modalities.

[Influence and Countermeasure of Respiratory Motion on Matchline Profile of Field Matching Technique in Passive Scattering Proton Therapy for Esophageal Cancer].

This study aimed to evaluate the influence of change in respiratory motion on matchline (ML) and reduction of the effect by increasing ML levels of field matching technique in passive scattering proton therapy for esophageal cancer. To evaluate the influence of respiratory motion in terms of stability, we measured relative dose around ML using a respiratory motion phantom. The relative error was -0.5% when the respiratory motion phantom worked stable, whereas there was obvious change that the relative error was -25.5% when the difference of amplitude between upper field and lower field was one side 3 mm on each cranially and caudally direction. In clinical case of the seven esophageal cancer patients simulated by the treatment planning system, assuming the difference of amplitude was 3 mm, the relative error of maximum (minimum) dose in clinical target volume around ML against the original treatment plan were 5.8±1.2% (-6.0±2.7%), 3.3±0.9% (-3.8±1.0%), and 2.4±0.5% (2.6±0.8%) on average (±SD) when ML levels were 2, 4, and 6, respectively. Increasing ML levels can reduce the influence of respiratory motion.

Effect of anatomical change on dose distribution during radiotherapy for maxillary sinus carcinoma: passive scattering proton therapy versus volumetric-modulated arc therapy.

OBJECTIVE:: Maxillary sinus carcinomas are anatomically situated next to many organs at risk (OARs), and anatomical change is often observed during radiotherapy. We analyzed the effect of anatomical change on dose distribution of passive scattering proton therapy (PSPT) and volumetric-modulated arc therapy (VMAT) for 20 patients. METHODS:: The first plans were generated based on the first CT images. The second CT images were acquired after 3 weeks, and the second plans were generated by copying the first plans to the second CT images. The effect of anatomical change was estimated by comparing both plans. RESULTS:: Target volume change was observed in all cases, however, the influence on dose coverage of clinical target volume tended to be small. Alternatively, the doses to almost all OARs were increased. In particular, the increase in the dose to brainstem (p < 0.001) and optic chiasm (p < 0.001) was significantly higher in the second PSPT plan than in the first PSPT plan. Although PSPT is sensitive to anatomical change, the dose to OARs remained significantly lower in PSPT plans than that in VMAT plans. CONCLUSION:: PSPT was confirmed to be more effective than VMAT even the effect of anatomical change was taken into account. Therefore, it is expected that the contralateral vision can be preserved reliably while optimal target coverage is provided. ADVANCES IN KNOWLEDGE:: PSPT allowed significant sparing of OARs even in the result of the second plans affected by the anatomical change. PSPT offers benefits over VMAT in reducing dose to several OARs.

Feasibility of CBCT-based proton dose calculation using a histogram-matching algorithm in proton beam therapy.

The aim of this study was to confirm On-Board Imager cone-beam computed tomography (CBCT) using the histogram-matching algorithm as a useful method for proton dose calculation. We studied one head and neck phantom, one pelvic phantom, and ten patients with head and neck cancer treated using intensity-modulated radiation therapy (IMRT) and proton beam therapy. We modified Hounsfield unit (HU) values of CBCT and generated two modified CBCTs (mCBCT-RR, mCBCT-DIR) using the histogram-matching algorithm: modified CBCT with rigid registration (mCBCT-RR) and that with deformable image registration (mCBCT-DIR). Rigid and deformable image registration were applied to match the CBCT to planning CT. To evaluate the accuracy of the proton dose calculation, we compared dose differences in the dosimetric parameters (D2% and D98%) for clinical target volume (CTV) and planning target volume (PTV). We also evaluated the accuracy of the dosimetric parameters (Dmean and D2%) for some organs at risk, and compared the proton ranges (PR) between planning CT (reference) and CBCT or mCBCTs, and the gamma passing rates of CBCT and mCBCTs. For patients, the average dose and PR differences of mCBCTs were smaller than those of CBCT. Additionally, the average gamma passing rates of mCBCTs were larger than those of CBCT (e.g., 94.1±3.5% in mCBCT-DIR vs. 87.8±7.4% in CBCT). We evaluated the accuracy of the proton dose calculation in CBCT and mCBCTs for two phantoms and ten patients. Our results showed that HU modification using the histogram-matching algorithm could improve the accuracy of the proton dose calculation.

Results of long-term monitoring of normal-tension glaucoma patients receiving medical therapy: results of an 18-year follow-up.

PURPOSE: To conduct a ≥15-year follow-up assessment of the visual field (VF) in normal-tension glaucoma (NTG) patients receiving medical therapy and to identify risk factors for VF progression. DESIGN: A retrospective clinical study. METHODS: Medical records of 78 eyes of 78 NTG patients monitored for ≥15 years were reviewed. VF progression was defined by a mean deviation (MD) deteriorated twice by 3.00 dB from baseline (MD criterion) and an annual decrease in the MD slope exceeding -0.5 dB/year (MD slope criterion). Logistic regression analysis was employed to identify risk factors for VF progression. RESULTS: The mean follow-up period was 18.3 years. The average intraocular pressure (IOP) before treatment was 15.1 ±1.9 mmHg and the average treated IOP was 13.5 ±1.5 mmHg with 2.0 medications. Forty-two eyes (53.8%) showed VF progression using the MD criterion and 15 eyes (19.2%) showed a negative MD slope less than -0.5 dB/year. Disc hemorrhage (DH) was observed in 30 eyes (38.5%). The mean VF progression rate was -0.38 ±0.30 dB/year in the DH group and -0.24 ±0.28 dB/year in the non-DH group (P = 0.012). Multiple logistic regression analysis identified DH [relative risk (RR) 4.28; P = 0.028] as a risk factor for VF progression using the MD criterion. DH (RR 8.77; P = 0.007) and IOP fluctuation during follow-up (RR 5.03; P = 0.048) were detected as risk factors using the MD slope criterion. CONCLUSIONS: DH and IOP fluctuation were associated with VF progression in NTG during long-term therapy.

Case of endogenous endophthalmitis caused by Klebsiella pneumoniae with magA and rmpA genes in an immunocompetent patient.

We report a case of endogenous endophthalmitis caused by Klebsiella pneumoniae in an immunocompetent patient. A 73-year-old man with acute epididymitis who had no history of diabetes mellitus developed endogenous endophthalmitis. The patient underwent anterior vitrectomy and intracapsular cataract extraction with intravitreal injections of both vancomycin and ceftazidime. After the surgery, he was treated with topical and intravenous antibiotics; however, the left eye perforated and was enucleated. Culture from vitreous biopsy specimens grew as K. pneumoniae, which was positive for both magA and rmpA. K. pneumoniae should be considered as a pathogen that can cause severe endogenous endophthalmitis in patients with urinary tract infection. The severity of the disease may be related to the virulence genes.

Assessment of macular function of glaucomatous eyes by multifocal electroretinograms.

PURPOSE: To determine whether significant correlations existed between the morphological and functional parameters of the macular region of eyes with open-angle glaucoma (OAG). METHODS: Forty eyes of 40 OAG patients were studied. The morphological parameters were obtained by optical coherence tomography (OCT), and the functional parameters were acquired by automated Humphrey Field Analyzer (HFA) and multifocal electroretinograms (mfERGs). All of the tests were performed within 6 months of each other. The retinal thickness was determined by OCT in the nine Early Treatment of Diabetic Retinopathy Study (ETDRS) sectors of the macula, the fovea, and the four quadrants of the inner and an outer ring. The amplitudes of the second-order kernel responses of the mfERGs in the central 5° including the amplitude ratio of the nasal to temporal hemispheres (N/T amplitude ratio) were analyzed. The total mean deviation of the HFA corresponding to each OCT region was measured. The correlation between the different parameters was determined by coefficients of correlation and linear regression analyses. RESULTS: The N/T amplitude ratio of the second-order kernel responses of the mfERGs was significantly correlated with the retinal thickness in the inferior quadrant (r = -0.44; P = 0.004). There was a significant correlation between the N/T amplitude ratio and the threshold in the superior quadrant measured by the HFA Central 10-2 program (r = -0.40; P = 0.011) and also between the N/T amplitude ratio and the total deviation in the superior quadrant (r = -0.40; P = 0.010). There were significant correlations between the inferior retinal thickness and the average threshold and the TD in superior (r = 0.70, P < 0.001; r = 0.692, P < 0.001, respectively), nasal (r = 0.53, P < 0.001; r = 0.53, P < 0.001, respectively), and temporal (r = 0.46, P = 0.003; r = 0.44, P = 0.004, respectively) quadrants. CONCLUSIONS: Functional glaucomatous changes determined by mfERGs and perimetry are significantly correlated with the morphological changes determined by OCT.

Intraocular penetration of micafungin in patient with Candida albicans endophthalmitis.

PURPOSE: To investigate the penetration of micafungin, a new class of echinocandin antifungal agent, into the aqueous humor and vitreous after an intravenous administration. METHODS: Endogenous endophthalmitis caused by Candida albicans developed bilaterally in a 67-year-old man. Three hours before vitrectomy, the patient received an intravenous injection of 300 mg micafungin. Samples of aqueous and vitreous were collected during the vitrectomy approximately 60 min after the intravenous injection. The concentration of micafungin in both bodies was determined by high-performance liquid chromatography. RESULTS: The concentration of micafungin was 25.36 µg/mL in the serum, 0.026 µg/mL in the aqueous, and 0.043 µg/mL in the vitreous. The micafungin minimum inhibitory concentration (MIC) against the C. albicans strain isolated from our patient was 0.03 µg/mL. Thus, the micafungin reached the MIC in the vitreous. CONCLUSION: We suggest that intravenous micafungin should be considered in mild cases of endogenous fungal endophthalmitis, or be given in combination with other intravitreal antifungal agents with vitrectomy in more severe cases.

Case of endogenous endophthalmitis caused by Streptococcus equisimilis.

We report a rare case of endogenous endophthalmitis caused by Streptococcus equisimilis. A 74-year-old woman with endocarditis developed endogenous endophthalmitis. The patient underwent emergency mitral valvuloplasty, and intravitreal and subconjunctival injections of vancomycin and meropenem. After the surgery, she was treated with topical antibiotics, ointment, intravenous gentamicin and intravenous penicillin G potassium. The causative organism was identified as S. equisimilis. S. equisimilis should be considered as a pathogen that can cause severe endogenous endophthalmitis.

Case of endophthalmitis following intravitreal injections of bevacizumab.

PURPOSE: We report a case of infectious endophthalmitis that developed after the third intravitreal injection of bevacizumab. The endophthalmitis was caused by Staphylococcus epidermidis. METHODS: Observational case series. RESULTS: This was a case of a 77-year-old man who had received three intravitreal injections of bevacizumab and developed infectious endophthalmitis. Polymerase chain reaction of an aqueous humor sample showed that the endophthalmitis was caused by S. epidermidis. The patient received intravitreal ceftazidime and vancomycin, and subconjunctival injection of vancomycin. After the intravitreal injections he was treated with topical levofloxacin and cefmenoxime hydrochloride, and intravenous cefpirome. Subsequently, the ocular inflammation gradually decreased and his visual acuity remained at 30/200. CONCLUSIONS: Our findings indicate that endophthalmitis can develop after intravitreal bevacizumab injections. Polymerase chain reaction is useful for differentiation from sterile endophthalmitis.