Intraoperative high-dose rate of radioactive phosphorus 32 brachytherapy for diffuse recalcitrant conjunctival neoplasms: a retrospective case series and report of toxicity

IMPORTANCE

Adjunct treatments for conjunctival malignancies are needed when standard therapy provides limited benefits or fails.

OBJECTIVE

To describe the results of patients with diffuse conjunctival neoplasms treated with radioactive phosphorus 32 (32P)-impregnated flexible film.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective case series between January 1, 2010, and January 1, 2013, was conducted at Memorial Sloan-Kettering Cancer Center, a tertiary referral center. The study was conducted on 7 eyes of 6 patients treated for diffuse conjunctival squamous cell carcinoma, sebaceous carcinoma, or lymphoma that had recurrent or residual disease after primary treatment.

INTERVENTIONS

Patients underwent mapping biopsies and detailed conjunctival drawings to delineate the pathologic extent of the disease. The brachytherapy film used for treatment was the RIC Conformal Source Model 100 (RIC-100, RI Consultants). The RIC-100 is a flexible, thin (approximately 0.5-mm) film made of a polymer chemically bound to 32P. The radioactive 32P film was placed intraoperatively, allowed to stay in place until the prescription dose was reached, and then removed. The median dose at the prescription point (1 mm from the surface of the film) was 15 Gy (range, 5-17 Gy).

MAIN OUTCOMES AND MEASURES

Patients were tested for best-corrected visual acuity, recurrence-free survival, and adverse events scored by using the Adult Comorbidity Evaluation-27 scale.

RESULTS

Between 2010 and 2013, 7 eyes of 6 patients were treated. The median age of patients was 70 years. All patients had a recurrent or persistent neoplasm. Four patients with squamous cell carcinoma, 1 with sebaceous carcinoma, and 1 with metachronous bilateral lymphomas were treated. The median treatment time was 19 minutes (range, 10-52 minutes). The median follow-up was 24.9 months (range, 3.1-38.2 months). Recurrence-free survival 24 months after brachytherapy was 75% (95% CI, 19-89.1). Two moderate adverse events and 1 severe adverse event occurred. Visual acuity was stable or improved in 5 of the 7 eyes (ie, better than 20/70 in the 5 patients who retained their treated eye).

CONCLUSIONS AND RELEVANCE

Our results show the use of an intraoperative high-dose rate of 32P brachytherapy in selected cases of recalcitrant diffuse conjunctival neoplasms. This technique offers a novel adjunct in the treatment of these cancers. Further follow-up and study are warranted.

Qualitative evaluation of fiducial markers for radiotherapy imaging

PURPOSE

To evaluate visibility, artifacts, and distortions of various commercial markers in magnetic resonance imaging (MRI), computer tomography (CT), and ultrasound imaging used for radiotherapy planning and treatment guidance.

METHODS

We compare 2 solid gold markers, 4 gold coils, and 1 polymer marker from 3 vendors. Imaging modalities used were 3-T and 1.5-T GE MRIs, Siemens Sequoia 512 Ultrasound, Phillips Big Bore CT, Varian Trilogy linear accelerator (cone-beam CT [CBCT], on-board imager kilovoltage [OBI-kV], electronic portal imaging device megavoltage [EPID-MV]), and Medtronic O-ARM CBCT. Markers were imaged in a 30 × 30 × 10 cm(3) custom bolus phantom. In one experiment, Surgilube was used around the markers to reduce air gaps. Images were saved in Digital Imaging and Communications in Medicine (DICOM) format and analyzed using an in-house software. Profiles across the markers were used for objective comparison of the markers' signals. The visibility and artifacts/distortions produced by each marker were assessed qualitatively and quantitatively.

RESULTS

All markers are visible in CT, CBCT, OBI-kV, and ultrasound. Gold markers below 0.75 mm in diameter are not visible in EPID-MV images. The larger the markers, the more CT and CBCT image artifacts there are, yet the degree of the artifact depends on scan parameters and the scanner itself. Visibility of gold coils of 0.75 mm diameter or larger is comparable across all imaging modalities studied. The polymer marker causes minimal artifacts in CT and CBCT but has poor visibility in EPID-MV. Gold coils of 0.5 mm exhibit poor visibility in MRI and EPID-MV due to their small size. Gold markers are more visible in 3-T T1 gradient-recalled echo than in 1.5-T T1 fast spin-echo, depending on the scan sequence. In this study, all markers are clearly visible on ultrasound.

CONCLUSION

All gold markers are visible in CT, CBCT, kV, and ultrasound; however, only the large diameter markers are visible in MV. When MR and EPID-MV imagers are used, the selection of fiducial markers is not straightforward. For hybrid kV/MV image-guided radiotherapy imaging, larger diameter markers are suggested. If using kV imaging alone, smaller sized markers may be used in smaller sized patients in order to reduce artifacts. Only larger diameter gold markers are visible across all imaging modalities.

(106)Ru plaque brachytherapy for uveal melanoma: factors associated with local tumor recurrence

PURPOSE

Plaque brachytherapy is a common form of treatment for uveal melanoma, and the Collaborative Ocular Melanoma Study (COMS) used (125)I. Recently, (106)Ru has been reintroduced for plaque brachytherapy in the United States. We reviewed our experience treating uveal melanoma with (106)Ru plaque brachytherapy using COMS planning techniques, hypothesizing that we would observe similar outcomes to those in the COMS.

METHODS AND MATERIALS

Medical records of patients undergoing (106)Ru plaque brachytherapy were reviewed retrospectively. Patient, tumor, and treatment characteristics were recorded. Outcomes including visual acuity, local tumor recurrence, salvage treatment, metastasis, and survival were recorded. Cox regression analyses were used to determine factors associated with local tumor recurrence and enucleation.

RESULTS

Twenty-eight patients were studied. Median age was 60 years, and 50% were men. Median tumor base diameter and height were 9.4 and 2.6 mm, respectively. Ophthalmic complications were rare. Local tumor recurrence and enucleation occurred in 13 and 4 patients, respectively. Local tumor recurrence was associated with low visual acuity in the tumor-bearing eye, posterior tumors, small plaque size, and difference in plaque-tumor diameter of <6 mm. Enucleation was associated with low visual acuity and posteriorly located tumor. Estimated 5-year rate of death and metastasis was 18.5% and 11.4%, respectively.

CONCLUSIONS

Among patients treated with (106)Ru plaque brachytherapy using COMS planning techniques, we found a greater than expected rate of local tumor recurrence. Planning (106)Ru plaque brachytherapy should be done carefully at centers that have previously used COMS protocols and (125)I.

A comparative dosimetric analysis of virtual stereotactic body radiotherapy to high-dose-rate monotherapy for intermediate-risk prostate cancer

PURPOSE

Stereotactic body radiotherapy (SBRT) is being used with increasing frequency as definitive treatment of early stage prostate cancer. Much of the justification for its adoption was derived from earlier clinical results using high-dose-rate (HDR) brachytherapy. We determine whether HDR's dosimetry can be achieved by virtual SBRT.

METHODS AND MATERIALS

Patients with intermediate-risk prostate cancer on a prospective trial evaluating the efficacy of HDR monotherapy treated to dose of 9.5Gy×4 fractions were used for this study. A total of 5 patients were used in this analysis. Virtual SBRT plans were developed to reproduce the planning target volume (PTV) HDR dose distributions. Both normal tissue- and PTV-prioritized plans were generated.

RESULTS

From the normal tissue-prioritized plan, HDR and virtual SBRT achieved similar PTV V100 (93.8% vs. 93.1%, p=0.20) and V150 (40.3% vs. 42.9%, p=0.69) coverage. However, the PTV V200 was not attainable with SBRT (15.2% vs. 0.0%, p<0.001). The rectal Dmax was significantly lower with HDR (94.2% vs. 99.42%, p=0.05). The rectal D2 cc was also lower (60.8% vs. 71.1%, p=0.07). Difference in D1 cc urethral dose was not significantly different (87.7% vs. 75.2%, p=0.33). Comparing the PTV-prioritized plans, the rectal Dmax (94.2% vs. 111.1%, p=0.05) and mean dose (27.1% vs. 33.3%, p=0.03) were significantly higher using SBRT, and the rectal D2 cc was higher using SBRT (60.8% vs. 81.8%, p=0.07).

CONCLUSIONS

HDR achieves significantly higher intraprostatic doses while achieving a lower maximum rectal dose compared with our virtual SBRT treatment planning. Future studies should compare clinical outcomes and toxicity between these modalities.

Intraoperative and percutaneous iridium-192 high-dose-rate brachytherapy for previously irradiated lesions of the spine

PURPOSE

Advances in stereotactic radiosurgery have improved local control of spine metastases, but local failure is still a problem and repeat irradiation is limited by normal tissue tolerance. A novel high-dose-rate (HDR) brachytherapy technique has been developed to treat these previously irradiated lesions.

METHODS AND MATERIALS

Five patients with progressive disease at previously irradiated sites in the spine who were not amenable to further external beam radiation were treated. Catheters were placed intraoperatively in 2 patients and percutaneously implanted in 3 patients with image-guided techniques. Conformal plans were generated to deliver dose to target tissues and spare critical structures. Patients received single-fraction treatment using HDR iridium-192 brachytherapy.

RESULTS

Median dose was 14 Gy (range, 12-18 Gy) with a median gross total volume D90 of 75% (range, 31-94%); spinal cord/cauda equina dose constraints were met. At a median followup of 9 months, no local progression of disease has been observed. Four patients had reduction in pain 1-4 weeks after treatment. No brachytherapy-related complications have been observed.

CONCLUSIONS

Intraoperative and percutaneous iridium-192 HDR spine brachytherapy techniques were not associated with complications or acute toxicity. There has been no local progression at treated sites, and most patients experienced reduction in cancer-related pain.

Endoluminal high-dose-rate brachytherapy for early stage and recurrent esophageal cancer in medically inoperable patients

PURPOSE

The management of superficial primary and recurrent esophageal cancer (EC) in medically inoperable patients is complex. Endoluminal high-dose-rate (HDR) brachytherapy has shown mixed results in terms of toxicity and local control. In this study, we examined the outcomes and toxicities in a set of patients with superficial primary and recurrent EC treated with a consistent HDR technique.

METHODS AND MATERIALS

Between 8/2008 and 7/2011, 14 patients were treated with HDR intraluminal brachytherapy, 10 (71.4%) with recurrent disease, and 4 (28.6%) with previously unirradiated lesions. Patients received three weekly fractions to a median dose of 12 Gy (range, 10-15 Gy); dose was prescribed to 7-mm median depth with mucosal dose limited to 8-10 Gy using a 12-14-mm applicator.

RESULTS

Median followup was 15.4 months. Overall freedom from failure (OFFF) and overall survival (OS) at 18 months were 30.8% (95% confidence interval [CI]: 5.2, 56.4) and 72.7% (95% CI: 45.3, 100), respectively. For patients with recurrent disease, OFFF and OS at 18 months were 11.1% (95% CI: 0, 32.1) and 55.6% (95% CI: 15.4, 95.8), respectively. For patients with previously unirradiated disease, OFFF and OS at 18 months were 75.0% (95% CI: 31.6, 100) and 100.0%, respectively. Eight (57.1%) patients had Grade 1 acute adverse effects; 6 (42.9%) patients had chronic Grade 1 adverse effects; 1 (7.1%) patient developed Grade 2 stricture. Grade 3 tracheoesophageal fistula occurred in 1 (7.1%) patient. One patient died before completion of treatment of unrelated causes.

CONCLUSIONS

HDR endoluminal brachytherapy is a well-tolerated treatment for superficial primary and recurrent EC in medically inoperable patients.

Intraoperative 32P High-Dose Rate Brachytherapy of the Dura for Recurrent Primary and Metastatic Intracranial and Spinal Tumors

BACKGROUND:

Treatment of spinal and intracranial tumors with dural involvement is complicated by radiation tolerance of sensitive structures, especially in the setting of previous treatment.

OBJECTIVE:

To evaluate whether intraoperative brachytherapy with short-range sources allows therapeutic dose delivery without damaging sensitive structures.

METHODS:

The median doses of previous treatment were 3000 cGy (range, 1800-7200 cGy) for 8 patients with primary/recurrent and 17 patients with metastatic spinal tumors and 5040 cGy (range, 1300-6040 cGy) for 5 patients with locally recurrent and 2 patients with metastatic intracranial tumors. Patients underwent gross total or maximal resection of the tumor and were then treated with an intraoperative brachytherapy plaque consisting of a flexible silicone film incorporating 32P. A dose of 1000 cGy was delivered to a depth of 1 mm; the percent depth dose was less than 1% at 4 mm from the prescription depth. Median postoperative radiation doses of 2700 cGy (range, 1800-3000 cGy) were delivered to 15 spinal tumor patients and 3000 cGy (range, 1800-3000 cGy) to 3 intracranial tumor patients. The median follow-up was 4.4 months (range, 2.6-23.3 months) for spinal tumor patients and 5.3 months (range, 0.7-16.2) for intracranial tumor patients.

RESULTS:

At 6-month follow-up, for all spinal tumor patients, local progression-free survival and overall survival rates were both 83.3% (95% confidence interval [CI]: 62.3%-94.3%); for all intracranial tumor patients, the local progression-free survival rate was 62.5% (95% CI: 23.8%-90.9%) and the overall survival rate was 66.7% (95% CI: 26.7%-92.9%). There were no intraoperative or postoperative complications secondary to radiotherapy.

CONCLUSION:

Use of the 32P brachytherapy plaque is technically simple and not associated with increased risk of complications, even after multiple radiation courses. Local control rates were more than 80% in patients with proven radiation-resistant spinal disease.

American Brachytherapy Society consensus guidelines for high-dose-rate prostate brachytherapy

PURPOSE

A well-established body of literature supports the use of high-dose-rate (HDR) brachytherapy as definitive treatment for localized prostate cancer. Most of the articles describe HDR as a boost with adjuvant external beam radiation, but there is a growing experience with HDR monotherapy.

METHODS AND MATERIALS

The American Brachytherapy Society has convened a group of expert practitioners and physicists to develop guidelines for the use of HDR in the management of prostate cancer. This involved an extensive literature review and input from an expert panel.

RESULTS

Despite a wide variation in doses and fractionation reported, HDR brachytherapy provides biochemical control rates of 85-100%, 81-100%, and 43-93% for low-, intermediate-, and high-risk prostate cancers, respectively. Severe toxicity is rare, with most authors reporting less than 5% Grade 3 or higher toxicity. Careful attention to patient evaluation for appropriate patient selection, meticulous technique, treatment planning, and delivery are essential for successful treatment.

CONCLUSION

The clinical outcomes for HDR are excellent, with high rates of biochemical control, even for high-risk disease, with low morbidity. HDR monotherapy, both for primary treatment and salvage, are promising treatment modalities.

A comparison of the impact of isotope ((125)I vs. (103)Pd) on toxicity and biochemical outcome after interstitial brachytherapy and external beam radiation therapy for clinically localized prostate cancer

PURPOSE

To compare biochemical outcomes and morbidity associated with iodine-125 ((125)I) and palladium-103 ((103)Pd) brachytherapy as part of combined modality therapy for clinically localized prostate cancer.

METHODS AND MATERIALS

Between October 2002 and December 2008, 259 patients underwent prostate brachytherapy ((125)I prescription dose, 110Gy: n=199; (103)Pd prescription dose, 100Gy: n=60) followed by external beam radiotherapy (median dose, 50.4Gy). Eighty-seven patients also received neoadjuvant androgen deprivation therapy. Toxicities were recorded with CTCAE v 3.0, International Prostate Symptoms Score (IPSS), and International Index of Erectile Function questionnaires.

RESULTS

Overall, acute Grade ≥2 genitourinary toxicity occurred in 21% and 30% of patients treated with (125)I and (103)Pd, respectively (p=0.16). There were no significant differences in IPSS change or urinary quality-of-life scores between the isotopes at 4, 6, or 12 months (p=0.20, 0.21, and 1.0, respectively). IPSS resolution occurred at a median of 11 and 6 months for (125)I and (103)Pd patients, respectively (p=0.03). On multivariate analysis, only the use of neoadjuvant androgen deprivation therapy was predictive of time to IPSS resolution (p=0.046). Late Grade ≥2 gastrointestinal toxicity occurred in 7% of (125)I patients and 6% of patients treated with (103)Pd. Of 129 potent patients at baseline, there was better erectile function in patients who received (103)Pd (p=0.02); however, the followup was shorter for these patients. The 5-year prostate-specific antigen relapse-free survival for (125)I and (103)Pd patients was 95.2% and 98.2% (p=0.73), respectively.

CONCLUSION

There were no differences in acute or long-term genitourinary or gastrointestinal toxicity between (125)I and (103)Pd in combined modality therapy for prostate cancer. There may be less erectile toxicity with the use of (103)Pd; however, additional followup of these patients is needed. There was no significant difference in 5-year prostate-specific antigen relapse-free survival between (103)Pd and (125)I.

Intraoperative real-time planned conformal prostate brachytherapy: Post-implantation dosimetric outcome and clinical implications

PURPOSE

To report the dosimetric outcome of patients with clinically localized prostate cancer treated with I-125 permanent implantation using an intraoperative real-time conformal planning technique.

METHODS AND MATERIALS

Five hundred and sixty-two patients with prostate cancer were treated with I-125 permanent interstitial implantation using a transrectal ultrasound-guided approach. Real-time intraoperative treatment planning software that incorporates inverse planning optimization was used. Dose-volume constraints for this inverse-planning system included: prostate V100 >or=95%, maximal urethral dose <or=120%, and average rectal dose <80% of the prescription dose. Day zero computed tomography scans were acquired for post-implantation dosimetric evaluation.

RESULTS

The median V100 and D90 to the prostate target were 96% and 166 Gy, respectively. In 91% of cases a D90 of >or=140 Gy was achieved. In these patients, the V100 and D90 values did not have a significant influence on PSA relapse-free survival outcomes. The median maximum rectal dose and urethral doses were 104 Gy (72% of the prescription dose) and 187 Gy (130% of the prescription dose). The average and maximum rectal doses exceeding 100% of the prescription dose were less than 1% and 10% of patients, respectively. Average and maximum urethral doses exceeding 150% of the prescription dose were noted in 3% and 24% of patients, respectively. Average and maximum urethral doses exceeded 120% of the prescription dose in 21% and 58% of patients, respectively. Among patients where >or=2.5 cm(3) of the rectum was exposed to the prescription dose, the incidence of late grade 2 toxicity rectal toxicity was 9% compared to 4% for smaller volumes of the rectum exposed to similar doses (p=0.003). No dosimetric parameter in these patients with tight dose confines for the urethra influenced acute or late urinary toxicity.

CONCLUSION

Real-time intraoperative planning was associated with a 90% consistency of achieving the planned intraoperative dose constraints for target coverage and maintaining planned urethral and rectal constraints in a high percentage of implants. Rectal volumes of >or=2.5 cm(3) exposed to the prescription doses were associated with an increased incidence of grade 2 rectal bleeding. Further enhancements in imaging guidance for optimal seed deposition are needed to guarantee optimal dose distribution for all patients. Whether such improvements lead to further reduction in acute and late morbidities associated with therapy requires further study.

High-dose-rate interstitial brachytherapy in recurrent and previously irradiated head and neck cancers—Preliminary results

PURPOSE

Although high-dose-rate brachytherapy (HDRBT) offers significant advantages over low dose rate brachytherapy, there are scant data on improved local control (LC) and treatment-related complications in patients with recurrent head and neck (H&N) cancers. We report our preliminary results in patients with recurrent H&N cancers treated with interstitial HDRBT.

METHODS AND MATERIALS

Thirty patients with recurrent H&N cancers were treated with HDRBT between September 2003 and October 2005. Seventy-seven percent (23/30) of the patients had either local or regional recurrence in the area of previous external beam radiation therapy. The treatment sites were oral cavity/oropharynx (11/30), neck (10/30), face/nasal cavity (6/30), and parotid bed (3/30). Whereas 18 patients underwent surgical resection followed by HDRBT, 3 patients were treated with combined external beam radiation and HDRBT, and the remaining 9 were treated with HDRBT alone. The dose and fractionation schedules used were 3.4Gy twice per day (b.i.d.) to 34Gy for postoperative cases, 4Gy b.i.d. to 20Gy when combined with 40-50Gy external beam, and 4Gy b.i.d. to 40Gy for definitive treatment. HDRBT was initiated 5 days after catheter placement to allow for tissue healing.

RESULTS

With a median followup of 12 months, 6 local recurrences were observed 1-10 months after the procedure. The 2-year LC and overall survival outcomes for the entire group were 71% and 63%, respectively. Patients treated with surgical resection and HDRBT had an improved 2-year LC compared to the patients treated with HDRBT+/-external beam radiation alone (88% vs. 40%, p=0.05). Six Grade II and four Grade III complications were noted in five patients, all observed in the postoperative HDRBT group.

CONCLUSION

The preliminary results of HDRBT indicate an acceptable LC and morbidity in recurrent H&N cancers. A planned surgical resection followed by HDRBT is associated with improved tumor control in these high-risk patients. Based on these encouraging results, prospective clinical trials are warranted using HDRBT in recurrent H&N cancers to decrease late toxicity.

Five-year outcome of intraoperative conformal permanent I-125 interstitial implantation for patients with clinically localized prostate cancer

PURPOSE

To report the 5-year tumor control and toxicity outcomes for patients with localized prostate treated with I-125 permanent implantation using an intraoperative real-time conformal planning technique.

METHODS AND MATERIALS

Between January 1998 and June 2002, 367 patients with prostate cancer were treated with I-125 permanent interstitial implantation using a transrectal ultrasound-guided approach. Real-time intraoperative treatment planning which incorporated inverse planning optimization was used. The median follow-up time was 63 months.

RESULTS

The median V100 and D90 were 96% and 173 Gy, respectively. In 96% of cases a D90 of >140 Gy was achieved. The median urethral and rectal doses were 100% and 33% of the prescription doses, respectively. The 5-year PSA relapse-free survival outcomes for favorable and intermediate risk patients according to the ASTRO definition were 96% and 89%, respectively. In these patients no dosimetric parameter was identified which influenced the biochemical outcome. Of 38% who developed acute Grade 2 urinary symptoms, 63% had resolution of their symptoms within a median time of 6 months. The incidence of late rectal and urinary Grade 3 or higher toxicities were 1% and 4%, respectively. Seven percent (n = 27) developed late rectal bleeding (Grade 2) and 19% experienced late Grade 2 urinary symptoms.

CONCLUSION

Real-time intraoperative planning consistently achieved optimal coverage of the prostate with the prescription dose with concomitant low doses delivered to the urethra and rectum. Biochemical control outcomes were excellent at 5 years and late toxicity was unusual. These data demonstrate that real-time planning methods can consistently and reliably deliver the intended dose distribution to achieve an optimal therapeutic ratio between the target and normal tissue structures.

Calibration procedures for seeds preloaded in cartridges

Radioactive seeds preloaded in sterilized cartridges or needles are commonly obtainable from manufacturers. Under the US regulations for control of radioactive materials, seed users are required to account for all seeds and independently verify their air kerma strength (SK). As a result, the viability of inspection schemes that rely on measurement of aggregate seeds is of interest. In this paper we consider the conditions (if any) under which cartridge inspection can satisfy regulatory requirements and still provide practical benefit (i.e., time savings) against the regular single-seed assay. The standards for comparison are the recommendations of AAPM TG40, AAPM TG56, and ACR's "Standard for the Performance of Manually Loaded Brachytherapy Sources." The practical benefit is judged in comparison to the effort required to apply the 10% assay recommendation of TG40 to seeds in cartridges. Two specific cartridge inspection schemes are considered: (a) measuring the SK of each cartridge in a batch; (b) measuring a single cartridge sampled at random from the batch. Unlike the 10% assay, which is defined (imperfectly, in our view) without reference to the prevalence of in-calibration seeds, the estimation of the relative merits of cartridge inspection methods must necessarily include such information and, as such, is manufacturer specific. In this paper results are provided for Oncura model 6711 125I seeds in shielded and unshielded Mick cartridges. We show that the only practically useful cartridge inspection scheme is the batch scheme applied to unshielded cartridges. The false positive rates associated with the other schemes are such that we expect to open a cartridge (and perform the 10% assay) at least 80% of the time. While anything less than 100% of the time is theoretically an improvement, this neglects the additional effort required to assay the cartridges.

Combining pattern discovery and discriminant analysis to predict gene co-regulation

MOTIVATION

Several pattern discovery methods have been proposed to detect over-represented motifs in upstream sequences of co-regulated genes, and are for example used to predict cis-acting elements from clusters of co-expressed genes. The clusters to be analyzed are often noisy, containing a mixture of co-regulated and non-co-regulated genes. We propose a method to discriminate co-regulated from non-co-regulated genes on the basis of counts of pattern occurrences in their non-coding sequences.

METHODS

String-based pattern discovery is combined with discriminant analysis to classify genes on the basis of putative regulatory motifs.

RESULTS

The approach is evaluated by comparing the significance of patterns detected in annotated regulons (positive control), random gene selections (negative control) and high-throughput regulons (noisy data) from the yeast Saccharomyces cerevisiae. The classification is evaluated on the annotated regulons, and the robustness and rejection power is assessed with mixtures of co-regulated and random genes.