Salvage whole pelvic radiation after cancerous pelvic lymph nodes have been found

Is it still worthwhile to attempt salvage radiation (SRT) after positive pelvic lymph nodes (PLN) have been pathologically detected (stage pN1)? Traditionally, patients with PLN dissection (PLND)-diagnosed pN1 prostate cancer have been considered to have incurable systemic disease. Therefore, they were either observed until distant metastases were identified or started on lifelong androgen deprivation. Retrospective studies of the benefit of salvage whole pelvic SRT for pN1 patients have been equivocal: Abdollah et al. and Rusthoven et al. showed a benefit to salvage RT, but Kaplan et al.showed no benefit.

(Update 9/25/2022) Fonteyn et al. reported the results of the first randomized clinical trial, the PROPER RCT. Unfortunately, the trial did not reach its recruitment targets (n=64) and follow-up was shorter than planned.

• All patients had 1-4 positive PLNs found via PLND, most had 1 positive PLN
• Most (75%) also had prostatectomy, 64% had positive surgical margins

Patients were randomized to 2 years of ADT and either:

• Whole pelvic RT (as high as common iliac LNs) (WPRT), or
• Prostate bed-only RT (boost dose was allowed for known sites of recurrence) (PORT)

After 3-years of follow-up:

• Biochemical recurrence-free survival was 92% for WPRT vs 79% for PORT
• Acute grade 2+ gastrointestinal (GI) toxicity was 45% for WPRT vs 15% for PORT
• Late-term grade 2+ gastrointestinal toxicity was 21% for WPRT vs 3% for PORT
• Acute grade 2+ urinary toxicity was 52% for WPRT vs 48% for PORT (no diff.)
• Late-term grade 2+ urinary toxicity was 41% for WPRT vs 42% for PORT (no diff)

GI toxicity was probably increased by the wide margins used (5 mm) and by the fact that radiation was given on top of surgically dissected tissue. Such extended PLND is rarely done outside of Europe. The SPPORT and POP-RT trials, which did not include previous PLND, found no increase in GI toxicity. Now that PSMA PET/CT is available for recurrent patients, toxicity will probably be lower. The new MRI-targeting linacs (Viewray MRIdian and Elekta Unity) can control for intestinal motion that causes toxicity. Also, there may be an opportunity for better oncological results if para-aortic lymph nodes are also treated (see this link). A STAMPEDE trial found that 2 years of abiraterone increased oncological results.There may also be an opportunity to enhance results with apalutamide.

In an analysis of the National Cancer Database of 7,791 prostatectomy patients (treated from 2003-2010) who were staged pN1 after PLND, Zareba et al. found that most (63%) were initially observed without treatment, and an additional 20% received androgen deprivation (ADT)-only within a year of diagnosis. Only 18% received SRT, most of those (72%) with adjuvant ADT. Those treated with whole pelvic SRT+ADT had worse disease characteristics than those who were observed only: higher Gleason score, higher stage, higher positive surgical margin rate, and greater number of positive lymph nodes.

After 5.9 years median follow-up on 3,680 patients:

• Treatment with whole pelvic SRT+ADT decreased 10-yr mortality by 31% compared to observation only, and by 35% compared to ADT-only.
• Treatment with ADT-only or SRT-only was not associated with an increase in survival

Touijer et al. reported on 1,388 pN1 patients treated at three top institutions: Memorial Sloan Kettering (MSK), the Mayo Clinic, and San Raffaele Hospital in Milan. The MSK cohort was primarily only observed, the Mayo cohort primarily received lifelong ADT-only, and the Milan cohort was primarily treated with whole pelvic SRT+ADT As in the Zareba study, SRT+ADT patients had worse disease characteristics.

After 5.8 years median follow-up:

• Treatment with whole pelvic SRT+ADT decreased 10-yr mortality by 59% compared to observation only, and by 54% compared to ADT-only.
• Those with worse disease characteristics benefited the most.
• Treatment with ADT-only was not associated with an increase in survival compared to observation, although prostate cancer-specific survival was increased.

Zareba and Touijer did not report the toxicity of the salvage treatment, but with improved external beam radiation techniques and scrupulous image guidance, toxicity has been improving.

Zareba and Touijer had very similar outcomes. Although they were both retrospective studies rather than prospective randomized trials, it should be noted that the selection bias that typically plagues retrospective studies favored those who did not receive SRT+ADT. In spite of their worse disease characteristics, the patients who received pelvic SRT+ADT survived longer.

Recently we saw a similar advantage to pelvic SRT+ADT even in men who were not diagnosed as stage pN1 with a PLND (see this link). Taken together, these studies indicate a marked survival advantage to treating the whole pelvic area in men with pathologically diagnosed high-risk prostate cancer post-prostatectomy. A previous study found that among men with pN1, the ten-year incidence of distant metastases was 35%, suggesting that spread may be confined to pelvic lymph nodes for some time. This creates a unique window of opportunity during which salvage treatment may still be curative.

We have also seen evidence that high-risk patients with imaging-detected positive lymph nodes benefited from whole pelvic radiation as primary therapy (see this link).

These studies also constitute better evidence than we currently have that whole pelvic radiation with ADT is a better idea than picking off lymph nodes one at a time (for which we have no evidence of survival benefit). As we have seen (see this link), our ability to detect all cancer-infected lymph nodes is poor.

There are several variables that the patient and doctor must decide upon, and for which there is no clear evidence: duration of adjuvant ADT, amount of radiation, and the pelvic lymph node field. Clinical trials show that at least 6 months of adjuvant ADT with SRT even without lymph node involvement increases oncological effectiveness, the optimal duration is unknown and may vary with disease characteristics (see this link). The amount of radiation to the pelvic lymph node field seems to be about 50 Gy in most cases, and the amount given simultaneously to the prostate bed will ideally be at least 70 Gy (see this link). The extent of the treated area has been questioned recently. Studies show that infected lymph nodes are often missed in the common iliac area (see this link). There will be variations due to individual anatomy and known bowel sensitivity.

(Update 4/25/21) A major Phase 3 randomized clinical trial (INNOVATE) in 141 locations will determine whether intensifying the hormonal side of the treatment with 2 years of Zytiga+Erleada+ADT has better outcomes than 2 years of ADT. They prefer detection with an Axumin scan but allow PSMA and C-11 Choline too. The whole pelvic radiation dose will be determined by the treating radiation oncologist.

We're still not very good at finding cancerous pelvic lymph nodes

(Updated)

Pelvic lymph node (PLN) detection is important because it is one of the first places prostate cancer travels to after leaving the prostate or prostate bed. Cancer cells in the interstitial fluid of the prostate drain out into sentinel LNs and then into many other LNs. The LNs act like filters, catching the errant cancer cells. Sometimes the white blood cells surround and destroy the cancer cells, but sometimes the cancer changes the white blood cells and lymph node tissue, creating a microenvironment that is more hospitable to cancer cells implanting themselves there and growing. It can take years for enough cancer cells to create such a hospitable habitat and grow to a size that can be detected with a scan. When it is detected there, it is called stage N1, and is called "locally advanced." In some cases, prostate cancer may still be cured if it is locally advanced. The standard ways of detecting cancerous pelvic lymph nodes (PLN) are via surgical removal or radiographic detection. Neither is very good.

CT Scan

The standard of care for detecting positive LNs is a pelvic CT scan with contrast. (Sometimes MRIs are used for this with no advantage other than billing for the hospital.) This is often done the same day as a bone scan for high risk patients. The CT detects the size of LNs, and suspicion of cancer is as follows:

• < 8 mm: not suspicious
• 8-11 mm: gray area
• ≥ 12 mm: suspicious

The problem with detection by size is that many small LNs may harbor cancer, and enlarged LNs may be enlarged due to infection or due to inflammatory processes in nearby cells. The problem with CT detection is that it's possible for a LN with cancer to be small, and the cytokines released by it to enlarge a nearby LN that does not bear any cancer. While a biopsy of an enlarged node may be difficult to perform, enlarged nodes that shrink with androgen deprivation is a sure sign that cancer was causing the enlargement.

Patients with fewer and smaller positive LNs have longer survival, so if the patient wants treatment for N1 prostate cancer, whether local or systemic therapy, it is best to use an alternative method of detection.

USPIO MRI

Ultra-small paramagnetic iron oxide (USPIO) particles accumulate in healthy LNs more than in cancerous LNs. The particles and their lack can be detected in LNs using MRI. Combidex (ferumoxtran) is a brand of USPIO that is now available for this purpose, but only at Radboud University in Nijmegen, The Netherlands. It can detect positive LNs with a diameter as small as 2 mm in some cases (see this link). It is better than a C-11 Choline PET scan, which has a size limit of 6 mm. (Update 2/7/23) Fortuin at al. reported on a new trial there comparing PSMA PET/CT to USPIO with MRI.  In the same 45 patients, all with high risk of cancerous LNs (before primary therapy or recurrent):

• PSMA PET/CT found 71 suspicious LNs in 25/45 (55%) patients -2.8 per patient
• USPIO w/ 3T MRI found 160 suspicious LNs in 33/45 (73%) patients - 4.8 per patient
• MRI only missed 19 of the PSMA-detected LNs, but found over twice as many, and upstaged ⅓ more patients

In 20 patients, they used a 7T MRI and found:

• 115 suspicious LNs in 17/20 (85%) patients - 6.8 per patient
• 91% were smaller than 5 mm (the PSMA PET/CT size limit)
• 63% were smaller than 3 mm (the 3T MRI size limit)

Most of the suspicious LNs were outside of the surgical pelvic LN dissection area.

The outstanding question is: what is the value of detecting every last cancerous LN?

PET/CT or PET/MRI

As we've seen, the currently best PET scan is the DCFPyL PET/CT, which has been approved in the US for high-risk and recurrent patients. Tumor to background ratios may be especially better than the Ga-68-PSMA PET/CT scans. DCFPyL detected 30% more positive LNs in the same patients. It has high specificity (95%) but low sensitivity (~ 40%). The Axumin (fluciclovine) PET/CT is less accurate (see this link). PET/MRIs, now available at a handful of US institutions will provide greater accuracy. Detection of small metastases (< 5 mm) is unproven even in the best of these scans.

Meredith et al. reported on 532 patients diagnosed with the Ga-68-PSMA  PET/CT after PSA recurrence following initial treatment with prostatectomy (425 patients) or RT (107 patients).

• Among those treated with primary prostatectomy, positive lymph nodes were detected in 68%.  
• Among those treated with primary RT, positive lymph nodes were detected in 40%.

(Update 11/14/17) Schmidt-Hegemann et al. reported on 129 patients diagnosed with the Ga-68-PSMA PET/CT:

• 20 patients were scanned before initial RT treatment
• 49 patients were scanned after PSA recurrence after prostatectomy
• 60 patients were scanned after PSA persistence after prostatectomy (PSA never became undetectable)

Positive pelvic lymph nodes were detected in:

• None in the pre-initial treatment group
• 16% in the PSA-recurrent group
• 33% in the PSA-persistent group
• Detection rates were about the same in patients with PSA< 0.05 ng/ml

Multiparametric MRI (mpMRI)

Multiparametric MRI is more specific than CT, but is no more sensitive at detecting positive LNs. In one study, only 57% were correctly staged with a DW-MRI.

Surgical pelvic lymph node dissection (PLND)

Surgical removal, or PLND, is usually performed at the same time as a prostatectomy. The surgeon looks for about 5-10 PLNs and removes them for pathological analysis. In the US, this isn't done routinely by most surgeons because it is usually negative (only about 5% of prostate cancer patients have PLN invasion when first diagnosed), there are often false negatives, and there are risks of lymphocele and lymphedema from it. There are two indicators that it may be advisable to perform a PLND:

1. Risk of PLN invasion is greater than 2% (or 2.6%) on a validated nomogram like this one based on PSA, Gleason score and stage, or,
2. Enlarged PLNs have been detected with CT or MRI

(Note: This recent nomogram based on European patients recommends ePLND when the risk of PLN invasion is at least 7%)

When cancer is found, sometimes wider removal of as many as 30 PLNs is performed, called extended PLND or ePLND. The hope is to find more infected LNs and remove them, all of them if one is lucky, but the ability to control cancer using ePLND is controversial and the subject of clinical trials. ePLND is difficult because LNs are nearly invisible, small, and difficult to find, obscured by more colorful tissue and sometimes hidden in the visceral fat. Unlike blood vessels, which branch out, lymph vessels are networked. ePLND yield may be increased by injecting a fluorescent liquid, called indocyanine green, into the prostate and letting it drain through the lymph vessels. Even so, this missed 24% of sentinel PLNs in one study. A magnetometer that finds iron oxide particles that accumulate in lymph nodes has been tried intraoperatively (see this link). Radiotracers that consist of a gamma emitter (Indium 111 or Technetium 99m) attached to a PSMA ligand have also been used intraoperatively for this purpose in some recurrent cases (see this link) . PET scans may be used to detect some of the larger nodes to be removed. ePLND is a more common practice in Europe than in the US.

Even the most thorough ePLND misses positive PLNs. In one recent study, almost a quarter of positive LNs would have been missed even if ePLND had been used. Metastases don't just stick in sentinel LNs (the first ones that drain from the prostate). This is unlike breast cancer, for example. Cancer may accumulate in a LN without being detectable in all the LNs upstream from it.

The definition of the PLN field  of whole pelvic radiation as defined by a consensus of radiation oncologists missed 44% of the positive LNs, in this study. A study of LN failures after whole pelvic radiation therapy found that more than half had a failure above the treated area.

Clearly, there is no imaging modality that will find all metastatic cells in the PLN area. Failure of either ePLND or whole pelvic radiation to adequately treat the pelvic LNs that are most likely to be positive is problematic. As the coverage/dissection area expands, so does the risk of side effects. Lymphedema and lymphocele may result from ePLND. Late-term damage to the upper bowel is a risk of increasing the radiation field (see this link).

Such risks must be balanced against the evidence for benefits of treatment. The success of pelvic radiation in various settings was discussed here, and early results from the STAMPEDE clinical trial among N1 patients are encouraging.