Tuesday, January 27, 2009

Alternative Surgical Approaches for Rectal Cancer

Dr. John Migaly - general surgeon at Duke talks about Surgical Approaches for Rectal Cancer.

Rectal Overview

40k cases in us, 8k deaths per year

local recurrence in 10-40%

sexual dysfunction is a risk with surgery and should be evaluated preopertatively

7% risk of synchronous carcinoma, 30% risk of synchronous adenomatous change, therefore colonoscopy recommended.


Standard Approaches

  • T1 - transanal excision can be considered in T1
  • T2 - radical resection +/- postop RT
  • T3 or N1, or greater - preop CTRT followed by radical resection



Due to this EUS is necessary for appropriate selection T stage accuracy 83%, N stage accuracy 75%, however, this is extremely operator dependent.

T stage is usually overstaged with EUS, usually with T2s being read as T3.

MRI with coil may be close to EUS, perhaps slightly better for N stage.  This is being be utilized in the UK in particular to direct treatment selection.  According to CW, LP and the speaker, this should be considered more in the US.



Reviewed Dr. Heald's advocacy of the TME technique. (TME - Total Mesorectal Excision)

Mesorectum - LN, tumor foci, local extension, site of ECE from nodes.  Bounded by Denonviller's Fascia.

Lateral resection margins are important for survival and local control.  <1mm margins - 40% LR, >1mm margins 15% LR.  Similarly overall survival is improved with <4mm margings (Cawthorn, lancet 1990).

Arbmann BJS 1996 - TME improved LR and OS.

Stockholm I-II - (Martling, Lancet 2000) trials of preop RT followed by surgery.  TME was taught to all surgeons participating in the TME project.  both TME and RT greatly improved LRC in cross trial comparison.  APR rates also dropped from 60% to 27% when TME adopted.


Dutch TME trial proved the improvement in LRC with RT preoperatively (5Gy x 5), compared to TME alone.  MRC trial also demonstrates this as well.


CW brings up the risk of leaks with TME and diverting colostomies.  Diversion results in leak rates of ~5%, which is an improvement from prior.  We may be detecting more radiographically evident leaks with the adoption of regular CT scans in the post op setting, though clinically these may never have been relavent.


TME needs to be standard, but penetrance in the community is lagging.  Experience matters (K. Ludwig paper out of Duke shows ~5 fold improvement in sphincter preservation when >10 rectal resections are performed/year, compared to <3).


Need 1-2 cm margin to perform LAR, otherwise APR necessary.



J-pouch or side to end are primary options for rectum reconstructions.  Transverse coloplasty another option.  Leaks will often occur at the most caudal end of a J pouch.  Straight colo-anal anastomosis has poor reservoir capabilities.  Note if 6-8cm of rectum left, J-pouch probably not necessary.

Lightning Review of RCTs (Randomized Controlled Trials)

  1. J-pouch vs straight coloanal RCT - 89 pts (Hallbook, Ann Surg, 1996): Improvement in QOL with J-pouch.
  2. J-pouch vs coloplasty RCT - More leaks
  3. J-pouch vs side to end RCT - No difference


Laparoscopic TME:

Meticulous technique necessary, visualization may be superior.  Instrumentation improving.  Trendelenburg positioning necessary as this reduces the amount of small bowel in the field.  Outcomes still coming in.  Leak rates reported from 5-20%.  Early reports from 2000 demonstrate a steep learning curve.  In modern series, LRC has been as expected.  Selection key - need a non-obese patient, a non-bulky tumor.  Cosmesis, recovery, and length of stay improved.


ACOSOG Z6051 RCT of open vs lap TME resection for T1-3N1 or T3 rectal tumors less than 12cm from verge.  Goal accrual 440.  To open soon.  Primary endpoint will be margin status.


Local Excision:

Goal is a minimally invasive procedure to perform a full thickness bowel wall resection with good longitudinal margins along the bowel wall.


Unfortunately I missed this bit of the talk, but I think the best data is from the CALGB 8984 prospective trial of local excision in T1 and T2 tumors.  Selection: tumors must be <=10cm from verge,  <= 4cm in size, <=40% of circumference.  T1 tumors were treated with LE alone, and T2 tumors received postoperative CTRT (54Gy with 5FU 500mg/m2 x 2 cycles).  Local recurrence was acceptable in T1 tumors 8% at 10years, however in T2 LRF was 18% at 10 years.  Therefore LE may be considered for T1, appropriately selected, but not routinely for T2.




Monday, January 26, 2009

Wilms tumor

MP presented a case of wilms tumor (no apostrophe now per Halperin).

 4yo girl with 2 weeks of an enlarging abdominal mass.  Plain film obtained and then CT scan obtained.  Heterogeneous mass ~10cm in size with a rim of enhancing tissues on the edge of the mass. (note that imaging utilizing ionizing radiation was used first in this case; normally ultrasound would be the first modality).

 Differential of a Abdominal Mass - Wilms, Neuroblastoma, Renal Cell, Clear Cell, ATRT (atypical teratoid rhabdoid tumor).  Rim of enhancing normal kidney is a big tip off.  Calcifications can be see in Neuroblastoma.

 W/U - U/S, urinalysis (hematuria, - but really to look for VMA, HVA from neuroblastoma), CT, CXR or CT chest, ?Biopsy only if bilateral or unresectable.  A biopsy automatically makes the patient stage III, and requires WART (whole abdomen RT)!

 Surgery is an initial question - IVC, renal vein invasion, intraperitoneal seeding, assesemtent of other kidney (consider partial nephrectomy).

 Stage I - limited to kidney

Stage II - surgically completely excised, with invasion beyond kidney, local flank spillage (now stage III)

Stage III - residual tumor in abdomen, +LN, peritoneal spillage (all biopsies are classified as stage III)

Stage IV - distant mets

Stage V - bilateral disease

Return to case.  Intraoperatively tumor was spilled into cavity.  LN negative.  Path demonstrated wilms tumor, margin negative, no anaplasia.

 Most common abdominal tumor - median age 3-4 years (unilateral).  4-8% are bilateral and tend to present earlier 2.5 years.

 11-13% have a syndrome

1. WAGR - 11p del, WT1, PAX6: aniridia. 

2. Beckwith-Widemann syndrome

3. Denys-Drash

 10% have anaplasia - poorer prognosis

 5y Overall survival usually on the order of 85-95%

 Treatment - US:  adjuvant treatments (NWTS) vs Europe - Neoadjuvant treatment (SIOP)

Advantages and disadvantages of both approaches are straightforward.  Definitive diagnosis with a higher risk of intrapertitoneal contaimination occurs with adjuvant treatment, while neoadjuvant approaches improves the chance of complete resection, with a chance of mistaken diagnosis (5% mistreated in SIOP trials).

 To cover the US randomized data:

 NWTS 1 (1969-1974)

1. RT necessary in group I : no RT needed if <2years.  No dose response beyond 20Gy

2. VCR vs AMD vs VCR + AMD in Group II, III : VCR+AMD better

3. preop VCR in group IV: not helpful

NWTS 2 (1974-1979)

1. group I - more VCR/AMD is not helpful

2. groups II-IV - Adriamycin adds to VCR/AMD (everyone gets RT)

also RT must start within 9 days of surgery

NWTS 3 (1979-1985)

1. stage I - shorter chemo? - shorter chemo OK

2. stage II - no rt vs 20Gy - no RT OK (also adriamycin not needed)

3. stage II - 10gy vs 20Gy - 10Gy OK (when adriamycin used)

4. stage IV - add Cyclophosphamide? - Cyclophosphamide didn’t help 

NWTS 4 (1985-1994)

no RT questions, all groups randomized to pulsed ChT (less toxic, equally efficacious)


question of further refined risk groups


Current RT Reccomendations:

Stage I - no RT

Stage II - no RT unless unfavorable histology (10.8 WART), or spillage (10.8 to flank)

Stage III - 10.8Gy to WART, 21.6Gy to Gross disease

Stage IV - 10.8 Gy to WART, 21.6Gy to Gross disease

Lung mets - 12Gy at 1.5Gy/day

In Europe, the series of SIOP trials have demonstrated that preop AMD+VCR x 4weeks - surgery - risk adapted adjuvant treatment is safe and effective.  Worth keeping in mind that AMD+VCR can be used neoadjuvantly if not initially resectable.   RT dose with high risk factors (stage III disease in NWTS system, stage IIN1 or III in SIOP system) is 15Gy.

Case Resolution.  Pt treated to Whole Abdomen to 10.8Gy.  Simmed with Fluoro to encompass whole abdomen from superiormost extent of diaphragmatic movement, to the bottom of the ischial rings.  Femoral head blocks placed.  Need to discuss future issues with infertility with parents/patient.


Friday, January 23, 2009

Pancreatic Cancer - Adjuvant Therapy

JP Presents on adjuvant treatment for pancreatic cancer.

Case presentation is of a 55yo man presenting with obstructive jaundice:  Staging reveals a 2.5cm mass mod diff adenocarcinoma at pancreatic head via diagnosed via biopsy at EUS and ERCP.  Whipple performed with multiple positive margins at neck, with 4/10 lymph nodes positive.  (pT3N1).  PET CT was negative for distant disease.

BC: Whipple in this scenario was probably attempted initially as Pt seemed to be node negative at initial staging.  If nodes were known prior to the patient may have been better served by CTRT upfront, with reevaluation for resection afterwards.

34,000 pancreatic ca per year, with 32,000 deaths per year.  2/3 are found in the head.  80% are unresectable at diagnosis.  Risk factors: Smoking, chronic pancreatitis, obesity, high fat diet.  3y OS with resection 17%, at best 6% unresected.

Whipple classically resected distal stomach, duodenum, pancreatic head, common bile duct above the cystic duct, and the gallbladder.  For tumor in the body or tail, a distal pancreatectomy can be performed, sparing the head.  Spleen may be taken with this procedure.  Recently, pylorus sparing whipples have been gaining popularity at Duke.

T1 <=2cm, T2 >2cm, T3 out of pancreas, T4 celiac axis/sma

N0 no nodes, N1 nodes 

Randomized Trials Examining Adjuvant Treatment.

GITSG 9173 (Kalser, Arch Surg, 1985): n=43 (goal of 100, very poor and slow accrual), resected with negative margins.  XRT (40 gy split course) + 5FU vs. Observation.  2y OS 43% CMT vs 18% obs (p=0.03), Median DFS 11mo vs 9mo (SS).

EORTC 40891 (Ann Surg 1999) ChTRT vs obs.  2 yr OS 51% CMT vs 41% Obs (p=0.208).  Pancreatic head tumors seemed to do slightly better with CMT but not significant.  Again split course RT to 40Gy (split), low 5FU dose.  Poor treatment compliance (>20% in treatment arm did not receive treatment).  Updated in 2007 (link above)

ESPAC-1 (Neoptolemos, Lancet 2001) n=541.  This is a complicated study, which really is three trials, A: 2x2 trial of chemo, and chemort,  B: CMT vs obs, or C: Chemo vs obs.  Investigators could determine which trial patients went on.  RT remains antiquated at 40Gy split course.  No significant difference in CMT vs observation with pooling of the various trials.  Chemotherapy in the pooled arms did have a benefit.  A better analysis was presented of just the 2x2 trial (Neoptolemos, NEJM, 2004) - which demonstrated a benefit to chemotherapy alone, and a detriment to CMT.  The overall results are also poor compared to other trials. 

CONKO-001 (Oettle JAMA 2007). n=386 resected Pancreatic Cancer.  Randomized to Gemcitabine x 6 cycles vs observation.  Primary endpoint was DFS.  5y DFS was 16.5% vs 5.5% (SS).  OS trended toward significance - BC comments that on update, OS has become positive.

RTOG 97-04 (Regine JAMA 2008).  n=451 (388 for head tumors).  After surgery, randomization was to 5FU - ChtRT - 5FU vs Gem - ChtRT - Gem.  Dose was 50.4Gy.  Powered for OS for all comers and for the pancreatic head subset.  3yOS 31% Gem arm vs 22% 5FU in head tumors (p=0.09), HR 0.82 (95% CI 0.65-1.03).

So this leaves us with a mix of less than perfect trials, none of which answer the central question of wether modern RT is beneficial in the setting of adjuvant gemcitabine.  Stay tuned for upcoming randomized trials opening...

Monday, January 19, 2009

Neoadjuvant Chemo vs ChemoRT for GE junctional cancers

This one comes in from PG via email in regards to the newly published Stahl trial examing RT vs CTRT prior to resection in GEJ cancers:

The stats are a bit shy of the goals, but this appears to be clinically significant nonetheless. Results are consistent with the relative benefits of each treatment strategy according to the meta-analyses on this topic, but now there is a RCT.

One could argue that they did not use 'optimal' chemo (ECF), and instead used a different chemo regimen (cisplatin, fluorouracil, leucovorin), but the MAGIC trial compared chemo-surgery vs. surgery alone, so there is no direct comparison between these chemo regimens.  So, based on this RCT and meta-analyses, preop chemoRT appears to be superior in selected patients for these tumors.  No adjuvant chemo for either arm.

Monday, January 12, 2009

Christopher Counter: Oncogene addition in Pancreatic Cancer

Dr. Christopher Counter (Duke Associate Professor Pharmacology, Associate Professor Radiation Oncology) - Oncogene addiction in Pancreatic cancer


Below is a summary of Dr. Counters talk. I have left out a fair amount of the experimental details; all results come from in vitro animal models.


Ras Intro

EGF couples with GEF which then binds to Ras-GDP -> Ras->GTP (ras is membrane bound)

Ras-GTP interacts with RalGEF, Raf, PI3K

GAP then inactivates Ras-GTP -> Ras-GDP

KRas is mutated in 90% of pancreatic cancer, to the constitutively active form

in about 1/3 of human cancers this is also mutated, and predicts for resistance to EGFR inhibitors (specifically cetuximab in colon)

Ras oncogene addiction - Chin, Nature 2000: in a mouse model - turning off ras activation led to involution of the tumor.


Thinking of Tumor Initiation and Maintenance, and how Ras fits in:

Created a similar mouse model to Chin 2000 (mouse with a constitutive ras switch)

tumor is initiated in a mouse with ras active.

ras is then inactivated and then investigations focused on the three secondary pathways downstream from ras:

·         Ral-GEF only activation led to involution

·         Raf only activation led to involution

·         PI3K only activation led to tumor maintenance

Therefore for maintenance, it seems that PI3K is necessary for tumor maintenance.

Tumor microenvironment seems to activate a fair amount of paracrine signaling to maintain the tumor.


IL-6 story:

Ras induces secretion of IL-6

IL-6 is also found in pancreatic patients serum (along with IL-8)

if IL-6 is inhibited, tumor generated in xenograft model where much smaller and slower growing.  there was also a decrease in the amount of CD31 staining (an endothelial stain, and a surrogate for angiogenesis).

IL-6 knockout mice also were resistant to Ras-mediated tumorgenesis (papilloma model)

IL-6 monoclonal antibodies are available and are shown in mouse model to delay tumor growth.  in 10 human pancreatic lines tested, 2 had a response.  

IL-6 inhibition also seems to reduce cachexia. 


PI3K/AKT/eNOS story:

AKT is a downstream target from PI3K, and similar to the above seems to be a key component of tumor maintenance.

AKT works on a large number of targets, including eNOS (endothelial nitric oxide synthase), which also seems to be a key component of tumor maintenance.

Phosporylation of eNOS is performed on a specific serine (S1177), which seems to be necessary for tumor maintenance.

eNOS knockout model:  resistant to ras mediated tumorigenesis.

L-NAME inhibits eNOS, which was developed for another clinical application.  In xenograft model, L-NAME led to tumor growth delay.  in 9 cell lines, activity was seen in 7 (in contrast to the IL-6 mab intervention). (L-NAME developed for septic shock treatment by GSK)


eNOS nitrosylates Wild type Ras and activates it

N, K, and H are all subtypes of Ras, K being the most important for human tumors

eNOS (activated by KRas) then activates NRas and HRas via nitrosylation.

HRas nitrosylation seems to be required for tumor growth.


Farnesyl transferase inhibitors were developed to target Ras.  To be active Ras needs to get to the plasma membrane, most often by gaining franesyl groups.  FTIs inhibited the ability of Ras to get to the membrane, but it turns out that KRas is able to escape this by a number of other methods.  FTIs do however work well on HRas.  Perhaps some combination with the previously mentioned investigations would be fruitful.

Obviously a lot of this is ‘in the pipeline’ work, but it may well be key to improving the overall dismal results with current standard of care treatment. 

Tuesday, January 6, 2009

Oncoanatomy - Larynx

TZ presented the resident session of the oncoanatomy: larynx module.  Here are some notes that I culled from the presentation and resulting discussion

Le UCSF n=315, retrospective analysis of T1 Larynx - fx size less than 2.0 Gy were associated with lower local control.  for fx size of 2.25 or greater local control was 94%.  Their standard is 2.25Gy fx to 63Gy.

Several other retrospective series from MDA, and Florida exist with similar findings of inferior results with fx size less than 2Gy..

DB makes the case that in the Duke experience daily fraction doses of 2Gy to 64Gy have led to excellent local control (~94%).  He also comments that due to prescribing to the 95% IDL, the doses may have been closed to 2.1Gy.

Japanese randomized trial:

180 patients with T1N0M0. Randomized to treatment arm A (radiation fraction size 2 Gy, n = 89) or B (2.25 Gy, n = 91). Total Dose was 60 Gy in 30 fxs for tumors ≤  2/3 of vocal cord) or 66 Gy in 33 fractions for tumors > 2/3 cord in Arm A and 56.25 Gy in 25 fractions within 5 weeks for minimal tumor or 63 Gy in 28 fractions within 5.6 weeks for larger than minimal tumors in Arm B. 5yr LC was 77% for Arm A and 92% for Arm B (p = 0.004). 5yr CSS were 97% and 100% (NS). No differences in toxicity.

NOTES for treatment planning:

The figure of 8 refers to the anterior aspect of the thyroid cartilage that is visible on fluoroscopic simulation.  The ‘thin’ part of the cartilage, or the ‘x’ point of the ‘8’  is where the vocal cords attach to the thyroid cartilage and.  This is therefore a useful landmark for simulation and port checks.

For arrangement, parallel opposed can get the job done, but CW comments that CC Wang treated with a four field LAO, RAO, and opposed lats for homogeneity.  DB also remarks that he has been angling his beams slightly more oblique anteriorly to improve posterior coverage.

In regards to wedges, DB also makes the case for using 15deg wedges more than 30deg as he finds that there tends to be a cold spot in  the anterior commisure if the fields are over wedged. 

for T2, retrospective MDACC 180 pt series with improved LC with BID treatment. 

RTOG 95-12 ongoing T2a/b randomized to 70Gy 2Gy/day, HFX 79.2 1.2Gy BID.  Currently no significant differences in arms per report at ASTRO 2006. by Trotti.  5yr LC 70% QD vs 79% BID (p=0.11).  The magnitude of the difference is similar to what is seen in other HFX trials, so DB argues that this is an underpowered study that is probably showing a signal of efficacy.  

8.5Gy X 2 for Palliation of NSCLC

In chart rounds today, a question was brought up in re to palliation for lung disease.  A hypofractionated course of 8.5Gy x 2, separated by one week was prescribed and there were some questions about why this regimen was used.  

The best data supporting this is from a Randomized trial from France, published in the JCO by Sundstom in 2004.

421 patients with locally advanced stage III or stage IV NSCLC, with no curative options, with chest symptoms or centraltumor threatening the airways. 
Three arms: 
17 Gy / 2 fx 
42 Gy /15 fx 
50 Gy / 25 fx
There were no differences in survival or QOL measures.  

The caveats - in the 2 fraction arm, treatments were separated by 7 days.  They were also treated with steroids the day before, the day of, and the day after each fraction.  The length of spinal cord was limited to 12cm.  Also at three years there was a trend towards worse survival in the hypofractionated arm (3 yr OS 1% 2 fx, 9% 15 fx, 6% 25fx, p=0.06).

We have reserved this option for patients who's estimated survival is on the order of 3 months, and for whom transport for multiple fractions would be great hardship to them.