SpaceOAR hydrogel - is the difference worthwhile?

SpaceOAR hydrogel is spacer injected between the rectum and the prostate that increases their separation. This allows less radiation targeted at the prostate to hit the rectum. They published an interim 15-month report in 2015, and now Hamstra et al. have published their final report based on 3 years of follow-up.

The clinical trial was conducted at 20 participating institutions between 2012-2013 among patients who were to receive IGRT/IMRT as their first-line therapy for low- or intermediate-risk prostate cancer. Additionally, no more than 50% of biopsy cores were positive, ADT was not used, and prostates larger than 80 cc were excluded. The test was "single blinded:" physicians implanted the gel at the same time fiducials were implanted, but patients did not know if they received the spacer. All patients received 79.2 Gy of IMRT with 5-10 mm margins. 149 men were randomly assigned to the Spacer group. The Control group comprised 73 men.

While the researchers report physician-assessed toxicity data, they also collected patient-reported quality-of-life outcomes. Patient-reported outcomes are not subject to reporting bias, and are collected using well-validated questionnaires (EPIC). Patients did not know if they received the spacer. Patients filled out comprehensive questionnaires at baseline, 3 months after treatment, and then at month 6, 12, 15 and 36. On the EPIC questionnaires, a minimally important difference (MID) has been previously found to be a decline of 5 points, 6 points, and 11 points for rectal, urinary and sexual side effects, respectively. The researchers also reported the percent of patients whose quality-of-life scores declined significantly to a level that was at least twice as great (2X MID).

Acute toxicity

The change at 3 months compared to baseline represents the acute effects of radiation. This is the period of maximal deleterious effect of radiation on rectal and urinary quality of life. After 3 months, quality of life typically improves. The exception to this rule is sexual quality of life, which may continue to deteriorate, largely due to age (see this link).

At 3 months, the percent of patients who were bothered by any bowel-related side effect (moderate or big bother) was 9.4% among the Spacer group, and 5.7% among the Control group. The difference was not statistically significant. The only component (components included such morbidities of diarrhea, blood in stools, urgency, frequency, etc.) of bowel bother that was statistically significant was bowel pain, which was reported as a moderate or big bother by 6.8% of the Spacer group and none of the control group. The spacer made bowel pain worse rather than better.

Physician-reported grade 2 acute toxicity at 3 months was exactly the same (4%) for both the Spacer and the Control group. The spacer had no effect on any but the mildest acute toxicity.

Urinary scores were not significantly affected by the spacer. Among the Spacer group, 22.8% evaluated their urinary-related side effects as a moderate or big bother. Among the Control group, 17.1% evaluated it as a moderate or big bother. The difference between the two groups was not statistically significant, nor were any of the components of urinary function (e.g., pain, urgency, waking up to urinate, weak stream, frequency, etc.).

Bothersome sexual effects were also not significantly different between the Spacer and Control group, and at 3 months, were similar to baseline.

Late-term toxicity

At 36 months after treatment, 2.2% of the men in the Spacer group evaluated their bowel function as a big or moderate bother. This compares to 4.4% in the Control group -- not a statistically significant difference. None of the components of rectal bother were significantly better in the men who received the Spacer, although the scores were directionally better in almost every component. Quality of life in the rectal domain for both the Spacer and Control groups were close to their baseline values.

This low rate of bothersome rectal toxicity was confirmed by physician reports of rectal toxicity. Ignoring mild rectal side effects (i.e., grade 1, like blood spots on toilet paper or loose stool) that patients often do not bother to report to their doctors, we see that physician-reported grade 2 or higher rectal adverse events occurred in only 6% of patients, all in the Control group. Only one patient suffered a serious (grade 3) rectal injury in the Control group. It's worth noting that even this small number was an increase from the 15-month report. At 15 months, there was only one patient who had grade 2 or greater rectal toxicity. Late-term rectal toxicity is not a major issue with modern-day radiation.

Late-term urinary scores were not statistically significantly different for Spacer (4.4% big/moderate bother) and Control (8.9% big/moderate bother). Both were improved from baseline. The only component that was significantly worse was urinary frequency (18.2% vs 4.6%). It is unclear why any urinary side effects would be affected at all by a rectal spacer. It may be an artifact of the low sample size at 36 months - just 46 in the Control group, and 94 in the Spacer group.

There were no differences in overall sexual bother between Spacer and Control groups, or in any of its components. In the Control group, big/moderate sexual bother increased from 35.2% at baseline to 41.3% at 3 years. In the Spacer group, big/moderate sexual bother increased from 23.6% at baseline to 28.6% at 3 years. In both cases, sexual bother peaked at 1 year after treatment.

The following table shows the percent of patients in the Spacer and Control groups whose scores in each domain (rectal, urinary and sexual) declined by the minimally important difference (MID) or by at least twice the minimum (2x MID), which would be clinically significant to the patients.

*statistically significant difference between Control and Spacer with 95% confidence

We notice that the spacer moderated the declines in the rectal scores. There was a 16 percentage point improvement in late term rectal evaluations due to the spacer. However, we also saw that the relative decline was not very bothersome to the patients (96% did not find it even moderately bothersome).

What is perplexing here is why the urinary scores would have declined more without the spacer -the clinically notable difference (2x MID) was statistically significant.The dose received by the bladder and urethra was unaffected by the rectal spacer. There was also a rather sudden increase in both urinary and rectal MIDs at the 3-year mark. At all prior time points, there were no statistically significant differences between the Control and Spacer groups. Many respondents were lost to follow-up at 36 months (and over a third were lost) and data may become erratic as sample size gets small.

Is it worthwhile to treat all patients? 

The cost of the SpaceOAR hydrogel injection is about $2,500. IMRT patients should not expect any amelioration of bothersome acute rectal symptoms. A cost/value analysis depends upon which toxicity numbers one wants to focus upon.

1. Even if the difference were statistically significant (and it's not), only 2.2% would avoid bothersome late-term rectal symptoms by using the gel. So to spare one patient bothersome rectal symptoms, 45 patients would have to be treated at a cost of over $90,000. 
2. If we focus on the late-term toxicity improvement, 6% avoided late-term grade 2 or higher rectal symptoms by using a spacer. To spare one patient such symptoms, 17 patients would have to be treated at a cost of $42,500.
3. If we use the 2x MID difference as our guide, 16% would avoid low grade late-term rectal symptoms. So to spare one patient those symptoms, 6 patients would have to be treated at a cost of $15,000. 

Some would argue that even though the symptoms are relatively mild, late-term symptoms are often longer-lasting. Each patient has to decide for himself. Patients who are interested should obtain pre-approval from their insurance. It has been provisionally approved, but Medicare approval is expected in 2018.

Safety must be considered as well. The risks are not large, but neither are they non-existent. In addition to the dangers of anesthesia and infection, there is a small danger that the injection will penetrate the rectal wall or cause a rectal ulcer. Such dangers are small, and undoubtedly diminish with clinician experience.

Rather than treat all radiation patients, it would be much more efficient to identify the patient characteristics that increase the risk of late-term rectal toxicity. The presence of visceral fat, anatomic abnormalities, medications (e.g., anticoagulants and antihistamines), comorbidities (e.g., inflammatory bowel disease and diabetes), smoking, and the microbiome may play a role. There may also be a genetic component, like mutations in DNA repair genes, that renders a patient more susceptible to lasting damage from radiation. Further analysis may help explain why there was no benefit in terms of acute toxicity, and why even this small benefit took 3 years to show up. Dr Hamstra mentioned that analysis is ongoing, and they plan a follow-up presentation at ASTRO next year.

The researchers are to be congratulated on this very well executed randomized clinical trial. There are too few trials like this. In fact, Daniel Hamstra boasted:

"I think this is better than any other study ever performed for prostate cancer.  I know of no phase 3 trials testing a new technique or study which randomized patients and reported benefits for toxicity and QOL.  IMRT has not done it.  Image guided therapy has not.  Proton therapy has not.  Robotic surgery has not.  Nor has any surgical technique.  If you go back to the original Dutch 3D conformal trial published 15 years ago they showed reduced toxicity with 3D conformal RT as compared to 2D RT (but did not collect QOL).  So, this is really a landmark study in that it was randomized (and blinded in that the patients did not know which arm they were on until the end of the trial).

Note: Thanks to Daniel Hamstra for allowing me to read the full text, and responding to my questions.