Tuesday, December 30, 2008

Vulvar Cancer - GOG 37 and GOG 88

In re to conference this AM, there were some questions and confusion between the GOG 37 and 88 trials. Here are my notes on these two important trials -


Pelvic Nodal Control: GOG 37 (Homesley et al. OB Gyn 1986;68:733 PMID:3785783)

Resectable vulvar SCC

s/p radical vulvectomy & bil superficial and deep inguinal dissection found to have + inguinal LNs

114 pts randomized to:

Pelvic node dissection - PLND (15/53 patients had + pelvic LNs)

RT (45-50 Gy to inguinal nodes and pelvic nodes BUT NOT vulvar region)

50% of patients were cN0 (PE not sensitive for groin node mets)

Groin recurrence PLND 13/55 vs RT 3/59 (p=0.02)

OS (2) PLND 54% vs RT 68% (p=0.03)

Subgroup analysis showed benefit primarily to RT when >1 node was invovled, which has become standard for many gyn-oncs, however, this was a subgroup and should be interpreted with all of the standard caveats.


Inguinal Nodal Control: GOG 88(Stehman et al. IJROBP 1992;24:389)

Resectable vulvar SCC; Excluded T1 lesions unless LVI or >5mm invasion; s/p radical vulvectomy

58 pts randomized to

Bilateral groin dissection, RT if positive

Bilateral groin irradiation

50 Gy/ 25fx, with 50% of dose given with 12-13 MeV electrons; Rx'd to 3 cm

5/25 patients had + nodes in Arm A

OS and DFS were better in the dissected + selective RT arm, with a large portion of local failures in the RT arm in the groin.

However- the standard critique of this study is that RT was Rx'd to 3cm, which is clearly insufficient in many cases. See Koh IJROBP 1993;27:969-974 for a good discussion of this.


So there you go. GOG 37 is about pelvic nodes and GOG 88 is about inguinal.


CV mortality and androgen deprivation

Big day for RT and related studies in the JCO this week -

RTOG 85-31 - randomized trial in cT3, pT3 or pN1 patients with prostate cancer, randomized to RT alone vs RT + indefinite ADT. The final results showed a survival benefit to the RT + ADT arm.

This report is in regards to the potential cardiovascular effects of ADT, and thus patients were analyzed for CV mortality. No difference were seen among the arms, in fact the trend was in the wrong direction.

Caveat here is that these were high risk patients, and had a relatively high prostate cancer specific mortality (16 and 22% at 10 years) as a competing risk. This might not be applicable to lower risk disease.

Link

Quality of Life after PCI

Article from Slotman in the JCO this week:

This is a report from the PCI trial in extensive stage SCLC, looking at the quality of life in both arms. At 3 months, fatigue and hair loss were worse in the treatment arm, with a number of other scales slightly worse in the PCI arm as well.

Of course this is in light of the overall survival benefit of the treatment as reported in the NEJM. And one must also keep in mind selection bias in all of these QOL reports, in that if patients felt to ill to complete the form, no data is captured.

Still this is an interesting report and gives us data to base consent conversations upon.

Link

20 weeks too long to wait for RT after BCT?

Article in the JCO from Vancouver, finding worse outcomes in patients for whom RT was delayed beyond 20 weeks post surgery for breast conserving therapy (BCT). Note that this was in patients not recieving chemotherapy!

HR was near 2 for LR and breast cancer specific survival when RT start was delayed beyond 20 weeks.

Link

Induction Chemotherapy for Head and Neck Cancer

There is a good editorial in the JCO about induction chemotherapy for head and neck cancers, which sums up the arguments against adopting it as a standard of care prematurely.

Link

Friday, December 19, 2008

Rectal Cancer - Chemoradiotherapy

Rectal Cancer:

5y OS:
I 90-95%
II 75-95%
III 40-70%
IV 5%

Local Excision

RTOG 8902: Postoperative study. Low risk <3cm, T1, Gr 1-2, no LVSI, nl CEA. Treated with Observation after Local Excision 4% failure in T1. T2 and T3 were given CTRT and failure rates were 20-30%.

CALGB Local Excision update: Updated 2008
T1s were observed, T2 treated with postoperative CTRT after local excision. 10year LR for T1 were 8%; in T2 the rates were 18%, with a number of late local recurrences. This calls into question the ability to treat T2 tumors with local excision even with postoperative treatment.

MSKCC series (Paty, Ann Surg, 2002). Failure rates after LE were ~10% in T1, ~25% in T2 wether or not RT+/- chemotherapy was given.

T3-4:
Local Recurrence without further treatment beyond surgery was 25-40% in the pre-TME era. This is improved to 10-15% with TME technique.

Postop Data:
Oconnell NEJM 1994: Postoperative T3-T4 or N+ rectal cancer (n=666), randomized to RT with concurrent PVI vs bolus FU. Improved RFS and OS in the PVI group.

Smalley JCO 2006: Intergroup 0144 compared three different concurrent chemotherapy regimens. CI still looked better for LC and for heme toxicity (other arms were two different bolus regimens).

Preop Data:
Swedish Preop Trial (NEJM 1997):
n=1,168 T3+ or N+ treated in the preTME era. Randomized to preoperative RT 5Gy x 5 vs. observation. This is the only trial to show not only a LC benefit, but also an overall survival benefit to preoperative RT.

Dutch TME trial (Ann Surg, 2007)
n=1,861 all treated with TME, randomized to preoperative RT (5Gyx5) vs no neoadjuvant treatment. 5y LR 11% vs 6% (SS). No difference in overall survival. This is important in that it proved that RT was still needed for optimal local control, even after optimized surgery.

German Preop vs Postop Trial (NEJM 2004)
n=823, T3+ or N+. Randomized to preoperative CTRT (50.4Gy) with CI 5FU, or postoperative CTRT (55.8Gy) with CI 5FU. OS and DFS not different. 5 year Local Control 6 vs 13% (SS). Both acute and late toxicities were better with preoperative treatment. More sphincter preserving therapy in the preop group, in the subgroup that were determined not to be candidates prior to treatment (19 vs 39%, SS).

EORTC 22921 (NEJM 2006)
n=1,011. T3+, randomized to four arms, all recieved preop RT. The analysis is confused due to the multiple arms, but the take home is that chemotherapy is necessary, wether concurrent or postoperatively. LC was worse in the single arm that had no chemotherapy.

FFCD 9203 (JCO 2006)
T3-4 recieving preoperative RT then randomized to concurrent vs sequential chemotherapy. LC 17% vs 8% p=0.04. This proves the need for concurrent chemotherapy with Preoperative treatment.

Polish trial (B J Surg, 2006)
T3-4 n=2 randomized to 5Gyx5 or 50.4Gy with 5FU preoperatively. Higher toxicity in the CTRT arm. No difference in OS, DFS or LC. No difference in sphincter preservation. This trial is interesting in that it shows equivalence in a short course treatment and the more conventional prolonged CTRT approach. An additional article looking at late toxicities showed no difference between the arms (Rad Onc 2007).

Our approach-
EUS should be obtained in the majority of patients.

T1: Surgery alone. local excision can be considered an option, with careful consenting.

T2: Would reccomend more definitive TME surgery over local excision with postop CTRT.

T3-T4 or N+: Preoperative CTRT- WPRT to 45Gy in 1.8 Gy/fx, with concurrent CI 5FU or capecitabine, usually treated prone with a three field technique (PA, two laterals). An additional 5.4 Gy is delivered with lateral reduced fields, excluding small bowel, to the tumor and associated mesorectum/presacral region. Surgery with TME technique is then reccomended 6-8 weeks thereafter.

Thursday, December 18, 2008

RT in the AM better than in the PM?

File this one under odd randomized trials:

This Candian Trial randomized 216 patients undergoing RT for head and neck cancer were randomized to recieve RT in the morning (8-10AM) vs RT in the afternoon (4-6PM).  Why, you ask?  This is based on observations that oral mucosal cells exhibit a circadian rhythm tend to be in G1 in the morning and in M by the afternoon.  Thus RT should be less toxic to the mucosa in the AM.  

They didn't quite find statistical significance on their main endpoint, they did find some subgroup signals.  

http://www.redjournal.org/article/PIIS0360301608030356/abstract?rss=yes

Issues with Power and Multiple Randomized studies

In the discussion with regards to CTRT(chemoradiation) for cervical cancer, we got into a discussion of power and how this might explain the fact that 1 of the 5 studies examining CTRT vs RT was negative.

Definitions: Power is the chance that the study will find the difference as described in the statistical design. More formally, beta is the probability of accepting the null hypothesis, when in fact it is not true. This is a "type 2 error." Power is defined as 1-beta.

Most studies set this at 80% or that neighbor hood. Therefore, in a study with 80% power, the chance of not finding a real difference is 20%.

When you have 5 equivalent studies, all with 80% power, the chance of at least one of them commiting a type 2 error can be calculated as below:

1 - (0.8)^5 = 0.67 (this is most easily calculated by determining the chance that all of them won't make the error, and subtracting that from one).

Therefore from this thought experiment, we actually should not be surprised that one of these trials is negative, but rather that it was the most likely outcome.

Locally Advanced Cervical Cancer

AC presents a case of Locally Advanced Cervical Cancer.

Case: 42yo G4P3 woman presenting with hx of 3mo vaginal bleeding, and discharge. Biopsy positive for moderately differentiated SCC of the cervix. On exam, 5cm exophytic mass noted with parametrial involvement; cervix was fixed the sidewall. On PET, multiple Pelvic nodes were positive.

2nd most common cancer worldwide, with a predominance in south america, africa and south asia. In the US, it is the 6th most common (11,150 new cases in 2007), largely due to the widespread implementation of Pap testing.

Risk factors: HPV 16, 18 (E6 and E7 act on p53 and the Rb protein). Smoking, immunosuppression also associated. Prenatal DES is associated with clear cell adenocarcinoma.

Likely the incidence of cervical cancer will decrease with the increasing penetrance of the HPV vaccines.

Brief FIGO Staging Basics: IA microscopic, IB1 visible <=4cm, IB2 >4cm, IIA proximal vagina, IIB parametrium, IIIA distal 1/3 vagina, IIIB pelvic sidewall (including hydronephosis), IV involvement of bladder or rectum. Primarily a clinical staging system. Only exam and IVP may be used to stage, though hydronephrosis on CT can also be used to obtain IIIB. Stage IV diagnosis can only be arrived at after cysto or procto with biopsy proven involvement.

Of course, when determining treatment, one uses all of the information available including PET, CT, MRI and other modalities not included in the FIGO system.

1. Hreshcyshyn IJROBP 1979: IIIB, IVa, n=104, randomized to XRT+brachy +/- HU. Improved median survival 20 vs 11mos.

2. GOG85 IIB-IVa n=388 Whitney JCO 1999, XRT+brachy with HU vs 5FU/CDDP. All patients had operative nodal evaluations. PFS 57 vs 47%, OS 55 vs 43%, SS with one-tailed test.

3. RTOG 9001: Eifel JCO 2004 (update). IIB-IVA or IB-IIA >=5cm or positive nodes. EFRT (WPRT + PA nodes) vs WPRT with CDDP(75mg/m2)/5FU x 2, n=405. 8yr OS benefit: 67 vs 41% (SS), including local control, and DFS.

4. GOG 120: Rose JCO 2007 (update). IIB-IVA, n=526. XRT with CDDP (40 mg/m2) weekly, CDDP(50 mg/m2)/5FU/HU x 2, or HU twice weekly. Both CDDP arms were superior with 10yOS 53% vs 34% in HU arm, PFS and LC similarly improved (all SS).

5. GOG 109/Intergroup: Peters JCO 2000. IA2-IIA n=268, undergoing Rad Hyst and LND, with +LN, +parametrium, +margins randomized to RT vs RT+CDDP/5FU. 4yOS 71 vs 81%, with PFS benefit as well 63 vs 80%, all SS.

6. GOG 71: Keys 2003. IB2 (n=256) all underwent CTRT, randomized to observation vs TAH (doses of RT were slightly different between arms, 40Gy WPRT + 40Gy to pt A in the non-operative arm, 45Gy WPRT then 30Gy to pt A in the operative arm). 5y LC 27 vs 14% (no statistical significance). No PFS or OS advantage.

7. GOG 123: Keys NEJM 1999. IB2 (n=369), RT vs RT/CDDP then TAH. CDDP improved, OS (3yr 74 vs 83%), DFS, LC, all SS.

8. NCIC (Pearcey JCO 2002): n=253 IB-IVA. RT vs RT+CDDP (40mg/m2 weekly). 5yr OS 62 vs 58% (NS), no difference in PFS or LC. This is the outlier and reasons for why this trial is divergent remains a subject of debate.

Probably, however, it's simply due to a power issue as described during conference by DK and DB. I'm going to describe this in a separate post.

The approach at Duke is to treat with WPRT to 45Gy in 1.8Gy/fx, with weekly CDDP at 40mg/m2. This is followed by a brachytherapy boost via T&O to a total dose to point A of 75-85Gy. In this patient the PA nodes were treat in the initial fields (4field) and the boost was followed by a side wall boost of an additional 9 Gy.


Wednesday, December 17, 2008

3 articles on Lancet Online First

Three article of interest in this week's Lancet online:

1. Publication of the SPCG-7/SFUO-3 trial of androgen deprivation +/- radiation therapy in locally advanced prostate cancer. This trial was presented at ASTRO 2008, and was one of the highlights of the plenary session. 873 patients with largely T3 tumors were treated with indefinite ADT, and were subsequently randomized to RT or no further intervention. They demonstate a biochemical, cancer-specific and overall survival benefit.

While to believers this is a "tell us something we don't already know" moment, there are plenty of non-believers out there, and this trial fits into a much needed gap in the literature in the treatment of high-risk prostate cancer.

Then there is publication of two components of the MRC - ASTEC trial in Endometrial cancer-

2. The first is a question of lymphadenectomy. This is a randomized trial of TAH-BSO + washings and palpation of the PA nodes +/- Lymphadenectomy, with a sizable sample size (n=1408). Contrary to prior non-randomized reports (particularly out of Duke), this trial finds no benefit to lymphadenctomy beyond TAH-BSO. In fact, the HR is in the wrong direction, and was worse in the subsets where more nodes were harvested.

While there are significant differences with the surgical approach in the US and Europe, this trial seriously challenges this approach.

3. The second question is one of radiotherapy. This is actually a report of the ASTEC trial and a canadian trial (NCIC
CTG EN.5). Again a sizeable sample (n=905), randomized patients with intermediate to high risk, early stage endometrial cancer (any stage grade 3, IC any grade) to TAH-BSO to EBRT vs Observation. An important consideration is that a sizeable proportion (53%) received a brachytherapy boost to the vaginal cuff, outside of the randomization.

Not surprising, given findings from the PORTEC 1 and 2 trials, in addition to GOG 99, there was no survival difference. Local control was not formally evaluated in an actuarial fashion, however did seem to be improved. While there are no perfect studies examining this issued (at least from the rad onc's perspective), this again suggests that vaginal brachytherapy alone may be enough for the majority of early stage endometrial cancers.

Monday, December 15, 2008

CALGB 9633

Journal Club – CALGB 9633


A chemotherapy question today: what is the role of adjuvant chemotherapy in stage IB NSCLC?


Started with a case of NSCLC: 7x6cm mass in the LLL. On resection, negative margins, histopathology revealed a large cell carcinoma of neuroendocrine differentiation. All nodes were negative (hilar and mediastinum sampled).


Comment at this time about the neuroendocrine differentiation: CK states that this would be approached according to the NSCLC algorithm rather than as SCLC due to the large cell morphology.


IB survival rates on the order of 50-55%, with 30-40% long term relapse rates.


Review of prior modern adjuvant CT trials:

1. ALPI – scagliotti JNCI 2003.

n=1209 I-IIIA NSCLC randomized to MVP x 3 vs observation (majority received PORT).

OS – no difference


2. Big Lung Trial

n= 381 I-III NSCLC randomized to Cis/Vind, Mito/ifos/cis, Mito/vinb/cis, vinor/cis or observation, 14% received PORT.

OS – no difference


3. IALT

n=1867 I-III NSCLC randomized to cisplatin based chemotherapy vs observation, fair amount of PORT in this trial

OS 44% vs 40% at 5 years – statistically significant

update at ASCO: now negative as of ASCO 2008


4. JBR 10 (Winton NEJM 2005)

n=482 IB-II NSCLC Randomized to cis/vinor

OS HR 0.69 (0.52-0.91) SS (NS in IB subgroup analysis)


5. ANITA (Douillard, lancet oncol, 2006)

n=840 randomized to cisplatin based chemotherapy vs observations

5y OS 51% vs 43% (SS). Posthoc IB subgroup – no benefit.


6. JLCRG (Kato, NEJM, 2004)

n=999 stage I adenocarcinoma randomized to UFT vs observation.

5y OS 88% vs 85% (SS)


LACE meta-analysis (Pignon, JCO 2008)

pooled #1-5 above (~4500 pts). 5yr os benefit 5.4%


CALGB 9633

T2N0 lobe or pneumonectomy randomized to paclitaxel/carboplatin vs observation

powered for 13% OS benefit at 5 years. Accrual goal: 500pts.

80% power using a two tailed log rank (p=0.05)

well balanced amoung the arms

at the planned interim analysis(Strauss JCO 2004): 4 year OS showed a HR 0.62. stastical tests changed to a one-tailed test, and the accrual was stopped early.

At 5 yr however, the OS HR had become 0.82 and was no longer statistically significant.

In subgroup analysis tumors of >= 4cm, an OS benefit was seen.


Commentary - This is an overall negative trial with a hypothesis generating signal for a benefit in tumors >=4 cm in size. At the CALGB, statistical analysis would suggest that even if the accural went all the way to 500, the trial would still be underpowered for the observed HR. Another critique has been that carboplatin was used, which is likely slightly worse than CDDP in efficacy. Many have interpreted this data to mean that cisplatin should be used in the adjuvant setting, and that with a tumor >= 4cm, chemotherapy may be more strongly considered. This is the current practice at Duke.


Ongoing investigations

Intergroup trial: Investigating Chemo +/- Bev, including IB >=4cm though IIIA

The Muligene Model from Duke (Potti, NEJM) associated a gene expression signature with a high risk of recurrence in stage I NSCLC. A CALGB Study was to go forward with randomizing IB (2-4cm) in the high risk cohort to chemotherapy or not, though now revised so that everyone is getting randomized (low or high risk).


Case Resolution: Patient recived 4 cycles of cisplatin/navelbine, however has since recurred in the pleural space.

Friday, December 12, 2008

Tumor Lysis Syndrome

A Patient is presented with massive, untreated DLBCL, who experience tumor lysis syndrome after a single 2Gy fraction to a symptomatic R axillary mass.

Tumor Lysis Syndrome: An oncologic emergency caused by massive tumor cell lysis and release of large amounts of intracellular contents into the intravascular space.  Specifically results in elevation of serum Potassium, Nulcelic acids, and Phosphorous.

Uric acid can precipitate in the renal tubules, resulting in renal damage. 

Hande, et al. Am J Med 1993;94:133.  Study of high grade lymphoma treated with chemotherapy: 35-50% of patients had some increased in lab values after treatment.  Laboratory TLS was defined at a 25% increase in serum phosphate, K, or creatinine, or a decrease in calcium.  Pretreatment LDH and baseline renal function predicted for higher incidence of clinical tumor lysis syndrome.

Cairo, et al. Br J Haematol. 2004;127:3 - defined as a 25% change from normal values in creatine, calcium, phosphorus, BUN, or potasium.  Reviews multiple risk factors and treatments.

Coiffier, et al.  JCO 2008;26:2767 has a recent good review of TLS as well.

Interventions include hydration, rasiburicase (recombinant urate oxidase).  TLS may be prevention with allopurinol (inhibits the conversion of hypoxanthine and xanthine to urate).

 

Thursday, December 11, 2008

Hyperthermia in the Netherlands

Dr. Maarten Paulides from Erasmus in the Netherlands - an engineer associated with their Hyperthermia project - spoke about their current clinical efforts.

Background
12- year update of the netherlands deep ht trail
114 pts with locally advanced cervical cancer
1990-1996
ebrt + brachytherapy in all patients to 68-70Gy to point A
randomized to ht once weekly with heating goal of 42 degrees C for 60 minutes
local control 37% rt vs 56% rtht (p=0.001)
overall survival 20% rt vs 37% rtht (p=0.03)
actuarial > grade ii late toxicity was no different between arms

This is one of the strongest trials supporting a clinical benefit to hyperthermia with radiation therapy in the curative settting.  The primary weakness is that this trial occurred prior to the NCI alert in 1999 supporting the use of concurrent cisplatin based chemotherapy.

This weakness is addressed in an ongoing multi-institutional trial, currently open at Duke, exploring radiation therapy with concurrent cddp based chemo, plus or minus hyperthermia.

Current HT Clinic load at Erasmus:
deep ht - ~100 patients/year - 1.5 hour treatment time (0.5 hours of this is warm up), treated once weekly.
superficial ht - ~60 patients/year - 1 hour treatment times (0.25 hours for warmup), treated once weekly.

Ongoing ht trials at erasmus:
phase iii cervix trial (noted above)
recurrent rectal cancer: capcetabine + HT in previously irradiated patients
developing head and neck applicator

Head and Neck

Dr. Paulides referenced the below as backgound.
small randomized trial of RT +/- HT in locally advanced head and neck cancers
in 44 nodes (41 patients) the response rate was 41% rt vs 83% rtht
this trial also reported improve nodal control and overall survival at 5 years, though one must take the number of patients into account.

With this backdrop, the physicists and engineers at Erasmus have been developing a head and neck collar applicator for hyperthermia delivery, which has now been used for treating patients.  Monitoring is a bit of a problem as many of these sites are not accessible or amenable to invasive thermometry.  Addtitionally cooling from vasculature and respiration may confound attempts for homogeneous heating of the target.  However, he was able to show that this could be done, and I for one was quite surprised that they were implementing this in patients already, including a nasopharyngeal patient which he showed the thermometry for.  

Link to an Abstract describing the applicator (the Hypercollar): http://www.esho.info/esho2007/pdf/paulides1_esho2007.pdf

Tuesday, December 9, 2008

Vitamin E, C, and Selenium - Time to Curb your Enthusiasm?

Two very large randomized studies in early release from JAMA today, examining vitamin C, vitamin E and Selenium supplementation and cancer prevention.

Spoiler alert: All three had no effect on cancer prevention.

http://jama.ama-assn.org/cgi/content/full/2008.864

Cancer Stem Cells - that rare?

Nature has an interesting lead article this week on cancer stem cells.  By modifying the assay used to calculate the prevalence of tumorigenic cells, these folks from Ann Arbor have shown that they can increase the yield of human melanoma stem cells from 0.1 - 0.0001% to ~25%.  Subsequent testing revealed that CD133 in the high yield assay did not enrich for stem cells.

Significance: The assertion that tumorigenic cells are very rare and can be selected for by flow or other cell sorting methods is called into question, as the yield is so dependent on the model set up.  

Food for thought.

http://www.nature.com/nature/journal/v456/n7222/abs/nature07567.html

Recurrent GBM (Resident Report)

KC presented a case of recurrent GBM after concurrent RT/TMZ.

Primary Treatment Roundup:
1. Reviewed data from the Stupp Trial from the EORTC reviewing, improvement of 2yr OS from 8% to 26% (SS) with the addition of TMZ with RT in primary treatment after maximal safe resection.  

2. MGMT methylation status was reviewed, showing that methylation status was associated with improved survival with concurrent TMZ.  MGMT methylation silences the MGMT gene product, which repairs TMZ damage.

Duke approach to primary treatment: 
50.4 Gy 1.8Gy QD preop T2 with 2cm margin 
boost to 59.4 Gy primary postop T1 + contrast with 1.5cm margin.
all with concurrent TMZ and adjuvant TMZ afterwards.
most failures still occur within the treated field

Salvage Treatment Roundup:
1. TMZ was shown with multiple retrospective trials to not show great response.
Other cytotoxics (CCNU and derivative, Platinums) also do not show great responses.

2. EGFR inhibitors (TKIs) also under active review.  EGFR is amplified in ~50% of GBMs
JCO 22:133-142 gefitinib median EFS 8.1 wks, no radiographic response
PASCO 22L1502-55 EGFR TKI PR rate 6-25%
Imatinib + HU for PDGFR (Reardon JCO 2005): PFS @ 6mo 27%, radiographic response 9%

3. VEGF inhibition: promising due to high VEGF levels and angiogenesis in GBMs
JCO 2007 Vrendenburgh Phase II 35pt, irinotecan + bevacizumab (PMI17947719)
6mo OS 77%, 6mo PFS 46%, 57% PR, Median PFS 24wks, 1 yr OS 37%

CW comments that caution should be used in interpreting radiographic response with VEGF modulators given that bev can normalize vasculature and therefore reduce enhancement on MRI without necessarily having an effect on the actual tumor.

RT for Recurrent GBM
1.  Mayer IJROBP 2008: reirradiation of the brain.  
Necrosis at NTD>100Gy (normalize to 2Gy fractions)
Interval made little difference (though this is a limited study)

2. Combs JCO 2005
172 retrospective of FSRT to 36Gy in 2Gy fx, low rates of necrosis

3. Combs ACS 2005
SRS to 15 GY median OS 10mo after SRS
no Grade III or greater toxicity

4.  Cho IJROBP 1999
46treated with SRS 17Gy, 25 FSRT (37.5Gy/15fx)
11 mo SRS vs 12 mo FSRT
increased late complications in SRS group

5. Kong ACS 2008
prospective study of 114 pts treated with SRS after failure from RTCT
median 16Gy dose

6. ASTRO evidence based review 2005 -
no phase III data for salvage SRS
median survival after salvage SRS 6-12mo
historical control is ~7mo
no data on QOL
80-85% still die with local failure
for FSRT also no real difference in FSRT vs SRS detectable

Case resolution:  recurred radiographically at 20mo
15Gy with 5 dynamic conformal arcs, to the 100% isodose line
conformality index 1.8
maximum tumor dose 17.3Gy
Pt doing well at 1 month follow up

Attending comments: JK states that it is reasonable to treat a nodular recurrence of GBM with SRS.  Prospective, systematic study, with QOL measurements is clearly needed.  FSRT may be used in larger recurrences. Future directions may be with concurrent targeted therapies.  Also he comments that MD Anderson work suggests that bev after SRS may reduce adverse events secondary to edema after SRS (work from metastatic disease).